Why borewell data needs quality flags
Borewell readings are taken in difficult conditions: in heat and dust, by hand, from different reference points, often by several people across long shifts. Some readings will be wrong, some doubtful and some missing. The goal is not to make every dataset look perfect. It is to know which readings can support which decision.
A flag is a short label attached to a reading. It does not change the reading. It tells the next person how far to trust it and for what purpose. A water level that is slightly uncertain may be fine for a trend but not for a precise drawdown calculation.
Separate data quality from physical condition
A flag describes the reading, not the well. “Suspect reading” means the measurement is doubtful. It does not mean the aquifer is failing. Likewise, a sudden fall in compressor pressure might be a real event, such as a hose leak, or a gauge fault. Keep these two questions apart: is the measurement right, and if so, what does it tell us about the system?
Use a small, controlled vocabulary
Too many flags confuse people. A short list, applied consistently, works better.
| Flag | Meaning | Typical use |
|---|---|---|
| Good | Reading checked, no known problem | Use for all calculations |
| Uncertain | Plausible but with a known doubt | Use for trends; avoid for precise results |
| Corrected | Changed from the original with a recorded reason | Use with the correction record |
| Estimated | Not measured; filled by a stated method | Use only where the method is acceptable |
| Rejected | Unsuitable for the stated purpose | Keep, but exclude from calculations |
| Missing | No reading at the scheduled time | Show the gap; do not hide it |
Keep raw observations unchanged
The original reading must always survive. Corrections go in a separate field or record, never typed over the original. If a dipper reading was written as 31.8 m and later found to be 13.8 m because digits were swapped, keep 31.8 m in the raw column, add 13.8 m in the corrected column and record why. If the correction turns out to be wrong, the original is still there.
Automated checks should raise questions, not verdicts
Spreadsheet rules and logger software are good at spotting readings that need a look. They are not good at deciding what happened. Treat every automatic alert as a question for a person to answer.
- Range checks – a water level above ground or below the pump intake, or a compressor pressure above its rated maximum.
- Rate-of-change checks – a level that jumps several metres in a minute without a matching pump event.
- Sequence checks – timestamps out of order, or two readings at the same time with different values.
- Consistency checks – a flow meter total that does not match the sum of timed rates.
Manual review then adds context. A sudden level change may be real if the pump started at that moment. A flat line from a logger may mean a stable well or a stuck sensor.
What should a correction record contain?
- Which reading or readings were changed
- The original value and the corrected value
- The reason, with supporting evidence
- The method used for the correction
- Who made it and who checked it, with dates
- Which calculations and decisions may be affected
Common corrections on borewell sites
Measurement reference changes
Water levels are measured from a datum, usually the casing top. If the casing is cut, raised or fitted with a new head, every later reading shifts by that amount. Record the old and new datum heights and the date of change. Correct later readings to a common datum, or clearly report both series. A 0.4 m casing change can look like a 0.4 m water-level improvement if nobody notices.
Time corrections
Phone clocks, logger clocks and handwritten times drift or disagree. Correct timestamps only where there is clear evidence, such as a known pump start time recorded in two places. Record the offset applied. Be cautious: in early drawdown or recovery, a few minutes’ error can significantly change calculated results.
Interpolation versus observation
Filling a gap between two readings is estimation, not measurement. Flag interpolated values as estimated, record the method and never use them as if they were observed. Do not interpolate across a pump start or stop, where the level changes quickly.
Quality flags for compressor and drilling-air data
Drilling and air-lift records need the same care. Compressor discharge pressure, engine speed, hour meter readings and discharge temperature are often noted by hand during long shifts. Typical problems include gauges read at different points on the machine, pressure written in bar on one sheet and psig on another (1 bar ≈ 14.5 psi), and hour meter readings that go backwards because a figure was misread.
| Reading | Common error | Check |
|---|---|---|
| Discharge pressure | Mixed units or gauge location | State the gauge and unit on the sheet |
| Hour meter | Digits transposed, readings out of sequence | Hours must only increase |
| Discharge temperature | Taken at different points | Compare against normal oil-injected screw range of about 75–95 °C |
| Air-lift discharge | Timed bucket readings with frothy water | Allow the flow to settle; repeat timing |
Clean compressor data matters for service too. On hire jobs we change oil and service every 300 running hours, so an accurate hour meter log keeps maintenance on time.
Review the effect on decisions
A correction is only complete when its effect on results is checked. Rerun the affected calculations with the corrected data and compare with the earlier results. If a yield, drawdown or pump setting changes materially, the decision that used it must be reviewed.
Worked example: impact of a datum correction
A static level is recorded as 22.0 m below the original casing top. The casing is later raised by 0.5 m before pumping. The pumping level is read as 34.5 m from the new top. Uncorrected drawdown appears to be 12.5 m. Corrected to the same datum, the pumping level is 34.0 m below the original top, so drawdown is 12.0 m. At 9 m³/h, specific capacity changes from 0.72 to 0.75 m³/h per m. The difference is small here, but on a shallow drawdown it could be large.
Handle duplicates and impossible values
Where two people record the same reading, keep both and note any difference. Do not average incompatible values just to make a smoother number; the disagreement itself is evidence. For impossible values, such as a level below the bottom of the hole, flag as rejected, keep the original and try to find the cause in the field notes.
Review quality around pump and compressor events
Readings taken just before and after an intervention deserve extra attention. A pump start or stop, a change of discharge valve, a compressor being loaded or unloaded during air-lift, or a generator switching over can all produce rapid changes that are real but easy to misread. Note every intervention with its time in the field log, so a sudden change can be matched to its cause.
Power interruptions are common on rural sites, and each one interrupts a pumping test. Record the time of loss and return of power, and flag the readings that follow until the system settles. Do not join the data on either side of an outage as one continuous test without saying so.
Close the quality review with a disposition
At the end of a review, state what the dataset can support. For example: suitable for drawdown and recovery calculations from the second hour onwards; early-time data uncertain because of clock disagreement; flow meter total not verified. That statement is more useful than a general “data checked”.
How Reddy Compressors helps
We supply all types, all brands and all models of borewell compressors: new, used and refurbished, rentals, service and spare parts. We are an independent multi-brand specialist working with ELGi, Atlas Copco, Ingersoll Rand, KAESER, Chicago Pneumatic, Doosan and others, not an authorised dealer. We have tie-ups with all major and reputed brands.
When we supply compressors for drilling, development or air-lift work, we provide clear machine records: pressure, hour meter, service and any trips, in stated units, so your dataset starts clean. Our rental fleet covers 50–2,000 cfm and 2–30 bar. Groundwater interpretation remains with your hydrogeologist; we look after the compressed-air side. We are based at Tarnaka, Secunderabad, with over 20 service engineers and technicians covering Telangana and Andhra Pradesh, and we supply across India.
What to send us
- Compressor make, model and nameplate photo
- Your drilling or air-lift log, including pressures and hour meter readings
- Bit, rod and air-line details
- Any trips or faults recorded
- Site location and duration of work
Frequently asked questions
How do I correct a wrong water level reading in a pumping test?
Keep the original reading, add the corrected value in a separate column and write a correction record giving the reason, the evidence, the method and who checked it. Then rerun the calculations that used the reading and check whether any decision changes. Never overwrite the original, because the correction itself may later prove wrong.
What does a data quality flag mean in borewell testing?
It is a short label, such as good, uncertain, corrected, estimated, rejected or missing, attached to each reading. It does not change the value. It tells the reviewer how far the reading can be trusted and for which calculations, for example acceptable for a trend but not for a precise drawdown figure.
Can I fill gaps in pumping test data?
You can estimate missing values by a stated method such as interpolation, but flag them clearly as estimated and record the method. Never present them as observations, and do not interpolate across a pump start or stop, where levels change rapidly. Where a gap affects a key result, it is usually better to report it openly.
Why does my borewell water level look higher after casing work?
If the casing top was raised or cut, the measuring datum moved. Readings taken from the new top will differ from earlier ones by the change in height, even if the water level has not changed. Record both datum heights and the date, then correct all readings to a common datum before comparing them.
Why should compressor hour meter readings be checked?
Hour meter readings drive service intervals and hire records. Misread digits or out-of-sequence entries can lead to missed oil changes or disputed hire hours. Hours should only increase. A quick sequence check on the log, with a photograph of the meter at start and finish, prevents most problems.
Who should check borewell test data?
Ideally someone other than the person who recorded it, with knowledge of how the test was run. The checker reviews flags, corrections, datums, units and timestamps, then confirms what the dataset can support. For drilling and compressor logs, the site supervisor and the compressor supplier can confirm machine readings. Hydrogeological interpretation should rest with a competent consultant.
Related pages
Calculators and worked examples
These tools give indicative figures for screening, comparison and for preparing an enquiry. They run in your browser and send nothing. No prices or tariffs are assumed: enter your own to see costs. Efficiencies and other figures shown are example values: replace them with your own. Results are not a performance guarantee or a design certificate; final selection is confirmed by a service engineer against the machine data sheet and the site.
Drawdown with a measuring-point correction
Calculate drawdown from a static level and a series of pumping water levels, correcting readings taken after the measuring point changed (for example from casing top to ground level) and flagging rises during pumping.
The worked example below shows the method with fixed numbers. Turn on JavaScript to change the inputs.
Worked example
Inputs
- Levels recorded in: metres
- Static water level below the original measuring point (m or ft, as selected): 12.4
- Pumping water levels in time order, below the measuring point in use at the time (m or ft): 15.2, 16.1, 16.6, 16.9, 16.4, 16.5, 16.6
- Reading number from which levels were taken from ground level (0 = no change): 5
- Stick-up: height of the original measuring point above ground (m or ft; used only if a reading number above is entered): 0.6
- Rise between readings that counts as a step (m): 0.2
Result
- Correction applied: +0.6 m to readings 5 to 7 (ground level converted to the original measuring point)
- Final pumping water level: 17.2 m below the original measuring point
- Final drawdown: 4.8 m
- Maximum drawdown: 4.8 m
- Drawdowns after correction: 2.8, 3.7, 4.2, 4.5, 4.6, 4.7, 4.8 m
- Steps (rises) in the uncorrected data: at reading 5
- Steps remaining after correction: none
Method. All levels are depths below the original measuring point (casing top). A level read from ground level is converted by adding the stick-up (height of the casing top above ground). Drawdown = pumping water level − static water level, both from the same point. A drawdown that falls by more than the step threshold between two readings is flagged: during constant-rate pumping the level should not rise, so check for a pump stop, a flow change or a misread tape. Original readings are kept; the correction is shown separately.
What to take from it. A sudden shallower reading with no change in pumping is usually a changed measuring point or a pump stop, not the aquifer. Record the correction; do not delete the reading.
Technical content reviewed by D. Srinivas Reddy, Managing Partner and service engineer, Reddy Compressors. Last updated 17 October 2026. How our engineering content is reviewed.