Borehole Yield Test: What Your Well Can Really Give a Solar Pump

The short answer: a borehole yield test does not measure litres, it measures the water level while you take those litres. The two things you need out of it are specific capacity (yield divided by drawdown, in m3/h per metre) and the available drawdown (the distance from the static water level down to your pump intake). Multiply them and you have the maximum rate the well can sustain, and no pump, no panel count and no controller setting can push past it. A real 24-hour test in Muranga, Kenya held 8.1 m3/h over 70.4 m of drawdown; a second real test in Narok held 8.5 m3/h over 81.45 m. Almost the same litres. Sixty metres apart in pumping water level, which means one needs roughly 101 m of total head and the other about 162 m.

That gap is the whole reason this article exists. If you are buying a solar pump, the number that decides your model is not how much water the driller promised you.

What a borehole yield test actually measures

A yield test does not tell you how much water is underground. It tells you how much the water level falls when you take water at a given rate, and how quickly it comes back when you stop. Everything else is derived from those two curves.

Here is what that looks like on two real boreholes, both tested for a full 24 hours, both in Kenya:

Measured value	Muranga well	Narok well
Drilled depth	shallow volcanic borehole	250 m
Static water level	17.90 m	67.30 m
Pumping water level after 24 h	88.30 m	148.75 m
Drawdown	70.40 m	81.45 m
Average test yield	8.1 m3/h	8.5 m3/h
Specific capacity (Q divided by s)	0.115 m3/h per m	0.104 m3/h per m
Transmissivity from the test	2.76 m2/day	3.06 m2/day
Recovery in the first hour	4.14 m in 1 minute, 29.0 m in 60 minutes	60-minute recovery test
Head needed to push that yield to a tank 10 m up	about 101 m	about 162 m

Read the last two rows together and the trap is obvious. The two wells produce nearly the same litres per hour, and they need pumps that share almost nothing. The Muranga well will run happily on a 1100 W 110 V submersible. The Narok well sits at 148.75 m while pumping, and at that head you are into the high-head end of the range, where flow caps out around 4 to 5 m3/h in single-phase models.

The driller’s report usually quotes one number: the yield. Ask for the pumping water level as well. Without it the yield number is close to useless, because the same 8 m3/h costs you 60 m of head depending on which hole you drilled.

Specific capacity: the one number a yield figure hides

Specific capacity is simply the yield divided by the drawdown, in m3/h per metre. It is the honest description of a well, because it folds the size of the hole, the screen, the gravel pack and the aquifer into a single rate.

Source	Specific capacity (m3/h per metre)	Basis
Sandy aquifer, 10 tested boreholes, Mozambique	0.2 - 4.0	published test series
Hard rock, Karnataka, India	0.18 - 13.4	granite, basalt, limestone, laterite
Muranga, Kenya	0.115	8.1 m3/h over 70.4 m, 24 h test
Narok, Kenya	0.104	8.5 m3/h over 81.45 m, 24 h test

Two practical readings of that table. First, the same rock can differ by a factor of fifty depending on where the hammer opened a fissure. Second, both Kenyan wells tested below 0.12, and neither is broken. They are normal marginal boreholes, and a great many solar pumping projects in East and West Africa sit in exactly this band.

Specific capacity is also the single most useful number for a solar system, because it converts directly into a head penalty. When you take Q cubic metres per hour out of a well with specific capacity SC, the level drops by Q divided by SC metres, and your pump has to lift from that new level. So on a 0.115 well, moving from 2 m3/h to 4 m3/h does not just double the flow, it adds about 17 m of head:

Flow you ask for	Drawdown on a 0.115 well	Extra head your pump must supply
1 m3/h	8.7 m	8.7 m
2 m3/h	17.4 m	17.4 m
3 m3/h	26.1 m	26.1 m
4 m3/h	34.8 m	34.8 m
6 m3/h	52.2 m	52.2 m
8 m3/h	69.6 m	69.6 m

Note that the water level does not politely stay where you measured it. Every extra cubic metre per hour arrives with its own price in metres of head, and the price rises in step with the flow. That is the same coupling that makes a pump behave differently on a well than on a tank, and it is worth reading alongside how total dynamic head is calculated.

Why the second litre costs more than the first: s = BQ + CQ2

The line above is the idealised version. Real wells are worse than linear, and Jacob (1947) gave the reason: total drawdown has two parts.

s = BQ + CQ2

• BQ is aquifer loss. Water flowing through the formation toward the hole. Laminar, so it grows in proportion to the flow. You cannot change it; it is the geology you drilled into.
• CQ2 is well loss. Water entering the screen and travelling up the inside of the casing, where the velocity is high and the flow turns turbulent. It grows with the square of the flow.
• B and C are found by plotting s/Q against Q on plain arithmetic paper: the intercept is B, the slope is C. That is the Bierschenk (1964) graphical method, and it is the reason a step-drawdown test is worth doing.

Here is a real four-step test, published in the Muqdisho water supply groundwater study (NERC/NORA, 1980), with B and C fitted to the four points:

Step	Flow (m3/h)	Drawdown s (m)	s/Q	BQ (m)	CQ2 (m)	Well efficiency
1	19.58	8.79	0.4489	8.08	0.70	92.0%
2	30.38	14.20	0.4675	12.54	1.69	88.1%
3	39.25	19.05	0.4854	16.20	2.82	85.2%
4	49.67	25.00	0.5034	20.50	4.51	82.0%

Fitted result: B = 0.4128 m per (m3/h), C = 0.001828 m per (m3/h) squared, with an R2 of 0.9991. Divide the aquifer loss by the total drawdown and you get well efficiency, which fell from 92.0% to 82.0% as the rate rose from 19.6 to 49.7 m3/h.

That decline is the point. It is not noise and it is not measurement error; the turbulent term grows faster than the laminar one. The published guidance (Driscoll, 1986, as used in modern step-test reports) is that a well designed, constructed and developed properly should land at 70 to 80% well efficiency. A real step test on well APT-01 in Brunswick, Georgia shows how quickly that erodes with rate:

Flow (gpm)	Well efficiency
8	93.5%
17	78.0%
35	56.1%

Two consequences matter for a solar pump. First, the last cubic metre per hour is always the most expensive one in metres of head. Second, if the slope of your s/Q plot is steep, most of your drawdown is well loss, which means the well is under-developed or the screen is clogged, and that portion is potentially recoverable. If the slope is flat, the loss is in the formation and there is nothing to develop.

You can reproduce all of this arithmetic yourself. The graph paper version works, and it is also about fifteen lines of code; we keep the working in a script (scripts/well_yield_calc.py) so every number in this article can be re-derived rather than trusted.

Your real flow rate is where the pump curve crosses the well line

Here is the part that most sizing advice skips. A pump curve is measured with the pump pushing into a fixed head. A well is not a fixed head. So the flowing rate you get is not read off the pump curve at your calculated head. It is the point where two curves cross:

• the pump curve, descending: more flow means less head
• the well line, ascending: more flow means more drawdown, which means more head

Write the well line out and the reason for the scramble becomes clear:

TDH(Q) = static water level + Q divided by SC + lift above ground + pipe friction(Q)

Every term except the first grows with Q. Now put a real pump against it. Ours is the model in our own catalogue for which a single-model measured curve is published, so the points below are measurements and not a salesman’s shape:

Head (m)	88.94	87.2	79.96	67.01	52.0	42.53	37.63	28.25	18.66	5.3
Flow (m3/h)	0.58	1.06	1.56	2.02	2.54	3.11	3.51	4.04	4.51	4.95

That is 3DSC4.8-95-110-750-HV, a 750 W 3-inch stainless-impeller submersible on a wide-voltage controller (80-430 V). Now run it against four different wells, all with a tank 10 m above ground and 40 m of 32 mm PE delivery pipe:

Well	Flow with an unlimited aquifer	Real flow at the crossing	Total head at the crossing	Drawdown at that point
SC 3.0, static level 17.9 m	3.64 m3/h	3.64 m3/h	35.3 m	1.2 m
SC 1.0, static level 17.9 m	3.64 m3/h	3.53 m3/h	37.3 m	3.5 m
SC 0.115, static level 17.9 m	3.64 m3/h	2.51 m3/h	52.9 m	21.8 m
SC 0.104, static level 67.3 m	1.23 m3/h	0.99 m3/h	87.4 m	9.5 m

Same pump, same tank, same pipe. On the good well it is a 3.6 m3/h pump; on the marginal Muranga-type well it is a 2.5 m3/h pump, and the well has taken 31% of its output without ever being mentioned. On the deep marginal well it can barely produce a cubic metre an hour, because 87 m of head is almost the shut-off point of the model.

The crossing is found by iteration, and watching it settle is instructive:

Round	Assumed flow (m3/h)	Head the well demands (m)	Flow the pump gives at that head (m3/h)
1	3.00	58.4	2.32
2	2.32	50.8	2.61
3	2.61	54.0	2.47
4	2.47	52.5	2.52
5	2.52	53.1	2.50
6	2.50	52.8	2.51

It converges on 2.51 m3/h at 52.9 m of head. Note the loop is self-correcting: guess too high and the well pushes the head up until the pump backs off; guess too low and the pump speeds up until the well pushes back. That is worth reading next to how to read a pump curve, because the curve on the datasheet is only half of the picture.

The ceiling no pump can pass: specific capacity times available drawdown

There is one hard limit that has nothing to do with the pump:

maximum sustainable flow = specific capacity x available drawdown

Available drawdown is the vertical distance from the static water level down to the pump intake, less a safety margin of 3 to 5 m so the intake stays submerged and the dry-run protection does not trip. Set the pump 25 m below the static level and you have 25 m of available drawdown, give or take the margin.

Specific capacity (m3/h per m)	15 m available	25 m	40 m	60 m	80 m
0.104	1.56	2.60	4.16	6.24	8.32
0.115	1.73	2.88	4.60	6.90	9.20
0.30	4.50	7.50	12.00	18.00	24.00
1.00	15.00	25.00	40.00	60.00	80.00
3.00	45.00	75.00	120.00	180.00	240.00

All figures in m3/h. This table is the single most useful thing you can own about a marginal borehole, because it is not affected by which pump you buy, how many panels you hang or what the controller firmware is doing.

Now read it against the cost of depth. For a well with specific capacity 0.115 and a static level at 17.9 m, lifting the pump deeper raises the ceiling and raises the head at the same time:

Available drawdown	Ceiling (m3/h)	Head needed at that ceiling
15 m	1.73	44.5 m
25 m	2.88	57.0 m
40 m	4.60	77.4 m
60 m	6.90	107.9 m

There is the honest trade, in one table. Setting the pump deeper buys you water, and charges you head for it. Going from a 15 m setting to a 40 m setting roughly triples the ceiling, and adds 33 m of head, which takes you from a 750 W model to at least a 1500 W one. Neither half of that trade works alone: a deeper setting with the same pump just moves you to a worse point on the same curve, and a bigger pump over a shallow setting simply runs the well dry and trips.

Why a bigger pump does not fix a low-yield borehole

Buyers reach for more power, and the catalogue makes that mistake easy to make, because the same wattage is sold at wildly different duty points. All five of these are real models in our own range, all single-phase 110 V, and the last three all draw the same 1500 W:

Model	Power	Max. flow	Max. head	Behaviour on a well needing 53-77 m of head
4DSC25-26-110-1500	1500 W	25 m3/h	26 m	cannot lift from this well at all
4DSC20-48-110-1500	1500 W	20 m3/h	48 m	below the required head; no useful output
4DSC11-60-110-1500	1500 W	11 m3/h	60 m	works, with almost no margin above 53 m
4DSC7.5-100-110-1500	1500 W	7.5 m3/h	100 m	correctly matched
4DSC15-45-110-750	750 W	15 m3/h	45 m	not enough head, despite 15 m3/h on the label

Three of those share a power rating and only one of them will run your borehole. The two that fail do not fail because they are weak; they fail because they are low-head, high-flow machines, and a well with a deep pumping level is a high-head, low-flow job. On a marginal borehole, flow capacity on the label is not a feature. It is a warning that the pump was designed for a river, a pond or a cistern, where the head stays under 30 m all day.

The correct question is never “how many watts”, it is “what does the model deliver at my pumping water level, and does that exceed what the well can sustain?” If the first number is above the ceiling from the previous section, you have bought a pump that will cycle: it will pull the level down to the intake, the dry-run protection will cut out, the level will recover, it will start again. That pattern kills motors and it will not fill your tank any faster than a correctly sized smaller unit.

The recovery test: which half is the problem, the well or the aquifer

A step test tells you how the level falls. The recovery test tells you how it comes back, and it is the difference between two very different repair bills.

In the Muranga well, the pump was shut off after 24 hours at 8.1 m3/h. The level rose 4.14 m within one minute, and after 60 minutes it had risen to 59.30 m, recovering 29 m of the 70.40 m drawdown, or 41.2%. The shape of that is the information: an immediate jump, then a long tail. The fast part is the water stored right at the borehole and inside the screen coming back under its own pressure. The slow tail is the surrounding aquifer refilling the cone of depression, and it is governed by transmissivity, which in this well the test put at 2.76 m2/day. In the Narok well the same 60-minute recovery was reported as roughly 90%, and the report set the maximum safe design yield at 70% of the average tested yield, which is 5.95 m3/h out of the 8.5 m3/h tested.

Practical readings:

• Recovery is fast and near-complete in the same time it took to draw down: the aquifer is transmitting water well. Whatever your flow limit is, it is a stable limit and you can pump every day without depleting the hole.
• Recovery is fast at first and then crawls: normal for a marginal well, and it means the long-term rate should be set well below the tested rate. This is where the 70%-of-test figure comes from, and it is also what a storage tank is for: pump fewer hours at a lower rate, store the difference.
• The static level next morning is not where you left it: the well is being mined. Reduce the daily volume or add rest days, because no equipment change will fix a falling water table.
• The s/Q plot from the step test had a steep slope, so most of the drawdown is well loss: this is the one case where redevelopment is worth paying for. Surging, jetting or airlifting can clear fines and break up the mud cake left by drilling, and because it attacks the CQ2 term it can recover a genuine part of the flow. A well whose s/Q slope is flat has its loss in the formation, and redevelopment will not help it.

If the sand is also a problem, that is a separate conversation, and one we have written up in sand handling in solar boreholes.

How to run a step-drawdown test with the pump you already own

You do not need a hydrogeologist on site to get the two numbers that matter. You need a water level, a flow rate and a clock. The kit is a weighted tape or an electric water-level dipper, a gate valve on the discharge, a calibrated container, a stopwatch and a notebook.

1. Rest the well. Leave the pump off for at least as long as you are going to pump, and ideally overnight. Measure the level. That is your static water level, and everything depends on it being genuinely static.
2. Measure the flow volumetrically. Fill a 20 litre container and time it, three times, and take the average. Commercial test reports do exactly this; the Narok report required the discharge to vary by no more than 15% across a constant-rate test.
3. Step 1 at about one third of the rate you think the well can hold. Keep it steady for one to two hours, reading the level every 15 to 30 minutes. The step is finished when the level stops falling.
4. Step 2 at about two thirds, step 3 at full. Same duration, same readings. A fourth step at 1.2 times your intended duty is a useful safety check if the pump will allow it.
5. Compute s and s/Q for each step. Drawdown is pumping level minus static level. Plot s/Q against Q. Intercept is B, slope is C. If the slope is negative, you are inside the well-development problem described below and the numbers cannot be trusted yet.
6. Run the recovery test. Switch off and read the level at 1, 2, 5, 10, 15, 30 and 60 minutes. The early slope is the borehole; the late slope is the aquifer.
7. One absolute rule: never let the level fall below the pump intake during the test. If the intake starts sucking air you are measuring a pump, not a well.

A note on why the s/Q plot sometimes refuses to cooperate. Modern step-test guidance records that if the plotted points curve upward, the second-order assumption has failed and the greater-than-2 exponent form is needed. If they fit a straight line with zero or negative slope, the well-loss parameters are simply not measurable, and the usual causes are that the well is still being developed, that the test method was sloppy, or that the aquifer is contributing water unevenly along the screen. A negative slope is not good news about your well; it means the test did not produce an answer.

Safe yield: the two different two-thirds rules

This is where a lot of expensive mistakes get made, because there are three separate limits in circulation and they all sound like “about two thirds”.

Rule	What it limits	A real example	Source
Two thirds of the available drawdown	the rate a properly developed well may be tested and abstracted at	Sand Coulee Well No.5, Montana: available drawdown 162 ft, two thirds = 108 ft, potential yield 1,024 gpm. Its published loss coefficients reproduce 108.35 ft at 1,024 gpm	step and 24-hour test, 2012
70% of the tested yield	the ongoing operating rate, allowing for seasonal decline and imperfect recovery	Narok, Kenya: 8.5 m3/h tested, maximum safe design yield 5.95 m3/h	24-hour test pumping report
Permissible drawdown of 3 to 4.5 m	how far a suction-lift surface pump may pull the level down	a centrifugal pump on the surface cannot sustain more than about 6 to 7 m of suction	lift irrigation practice

All three are legitimate. They are also all different, and your system has to satisfy the strictest one that applies to it. If you are running a submersible, the suction-lift rule is irrelevant and you can use the full depth of the well. If you are running a surface pump beside the hole, the 3 to 4.5 m figure will always be your binding constraint long before the aquifer runs out, which is one of the reasons a riverbank or shallow-well job and a deep borehole job look nothing alike. For the deep-well case specifically, our 100 m borehole guide covers the voltage and cable side, which becomes the next constraint after the well itself.

Worked example: sizing a real Trista pump for a head-limited borehole

Take the Muranga-type well and size a solar system for it.

Input	Value	Where it comes from
Static water level	17.90 m	measured after 24 h rest
Specific capacity	0.115 m3/h per m	24-hour test, 8.1 m3/h over 70.4 m
Pump intake set at	62 m below ground	44.1 m below static level
Safety margin on the intake	4 m	keeps the intake submerged and the dry-run protection quiet
Available drawdown	40 m	44.1 minus 4
Tank height above ground	10 m	site layout
Delivery pipe	40 m of 32 mm PE	friction is computed, not guessed

Step one, the well’s ceiling: 0.115 x 40 = 4.60 m3/h. Nothing above that number is available from this borehole, ever.

Step two, the head at that ceiling: 17.90 + 40 + 10 + friction = 77.4 m. Note where that number came from. It is not the depth of the hole, and it is not the static water level. It is the static level plus the entire available drawdown plus the lift plus the pipe loss, and it is 60 m higher than anyone would guess from the well’s depth or from the driller’s yield figure.

Step three, choose a model that is comfortable above 77.4 m rather than bravely at it:

Model	Power	Voltage	Max. flow	Max. head	Fit for this well
3DPC3.8-95-48-750	750 W	48 V	3.5 m3/h	95 m	works, but caps you below the well’s own 4.60 m3/h ceiling
4DLR7.5-100-110-1100	1100 W	110 V	7.5 m3/h	100 m	good match, plastic impeller
4DSC7.5-100-110-1500	1500 W	110 V	7.5 m3/h	100 m	best match, stainless-impeller 4-inch
4DSC6-101-110-1100	1100 W	110 V	6 m3/h	101 m	good match, stainless impeller, less flow
3DSC5-135-110-1500	1500 W	110 V	5 m3/h	135 m	use this if the static level drops another 20 m in the dry season

The operating point of 4.60 m3/h at 77.4 m sits inside the envelope of the 100 m models with about a quarter of their head in reserve. That reserve is not decoration. Static water levels fall in dry seasons and over years, and the difference between a pump that survives a bad year and one that does not is usually those 20 m. The plastic-versus-stainless choice inside that ladder does not change the hydraulics at all, only how long the pump survives abrasive water; the trade is explained in plastic versus stainless steel impellers.

Step four, the daily volume: 4.60 m3/h times 5.5 effective solar hours is about 25 m3 per day, which is a serious village-scale supply rather than a garden. Design at 3.8 m3/h rather than at the ceiling and you still have 21 m3 a day with a genuine margin, at a head of about 62 m. That is the honest way to use a ceiling: aim below it.

Two electrical checks while you are there. Our catalogue specifies panel capacity at 1.3 times pump power or more, so a 1500 W pump wants at least 1950 W of array. And the array open-circuit voltage has to stay inside the controller’s limit: our 110 V pump models are specified for a VOC below 205 V, and the matching DF-110 controller is rated to under 220 V with an MPPT window of 110-150 V. Panels get colder than the air on a clear morning, and cold panels raise their voltage, so string for the cold-morning figure, not the nameplate one.

If your well turns out to be a head monster rather than a flow monster, the picture changes completely and the high-head end of the range is where you end up; our guide to sizing a solar water pump walks through that ladder, and pipe sizing is worth ten minutes of your time before you buy anything, because on a 100 m well the delivery pipe can quietly cost you as much head as the drawdown does.

The thing to take away from all of this is that a borehole has opinions. It decides how much water you get and it decides how hard your pump has to work, and it does both at the same time and in the same direction. Size the system around the well rather than around the yield figure in the driller’s report, and a marginal borehole becomes a perfectly ordinary engineering problem instead of a pump that keeps tripping. If you would rather have the arithmetic done for you, put your water level, your well’s specific capacity and your lift into the sizing tool, or send me your test figures directly.

---

Working out what your borehole can really deliver? Put your static water level, your specific capacity and your lift into the sizing tool to see your real head and your ceiling, or message me on WhatsApp with your static water level, pumping water level, tested yield, well diameter, lift and what the water is for, and I will come back with your specific capacity, the ceiling for your well, a model and a panel count that the well can actually support.