How a RINEX quality score is calculated
PosFlow combines completeness, multipath, cycle-slip and signal-strength indicators into a quality score. The weights and thresholds below are product settings, not an IGS acceptance standard. Use the score to prioritise review, then inspect the underlying metrics and station metadata.
Why a single number at all
A RINEX quality report has dozens of numbers in it, and most of them only matter in combination. A file with 99.8 % epoch completeness and 1.4 m of multipath is not a good file. A composite score exists so that a batch of two hundred sessions can be sorted, and the ones worth opening rise to the top. It is a triage tool, not a verdict: a score is a reason to look, never a reason to reject without looking.
The four components and what they weigh
PosFlow scores out of 100, distributed across four components. Anything an engine cannot report is excluded from both the earned points and the available total rather than scored as zero — otherwise a mode that does not measure signal-to-noise would be permanently capped below a mode that does, and the two would not be comparable.
| Component | Points | What it measures |
|---|---|---|
| Completeness | 40 | Observations actually recorded against those expected |
| Multipath | 25 | MP1 and MP2 linear combinations, 12.5 points each |
| Cycle slips | 20 | Slip rate per 1000 epochs |
| Signal-to-noise | 15 | Mean SNR on the first frequency |
The thresholds
Each indicator is graded Pass, Watch or Fail against fixed thresholds. These are the values PosFlow ships with:
| Indicator | Pass | Watch |
|---|---|---|
| Epoch completeness | ≥ 99 % | ≥ 95 % |
| Data completeness | ≥ 95 % | ≥ 90 % |
| Multipath MP1 / MP2 | ≤ 0.50 m | ≤ 0.65 m |
| Observations per cycle slip | ≥ 1000 | ≥ 400 |
| SNR (L1) | ≥ 38 dB-Hz | ≥ 32 dB-Hz |
Reading the grade bands
The composite score maps to five bands: 90 and above is Excellent, 75 Good, 60 Fair, 40 Poor, and below 40 Bad. The bands are wide on purpose. The difference between 82 and 86 is noise; the difference between 82 and 45 is a field visit.
Completeness is two different questions
Epoch completeness asks whether the receiver logged an epoch when it should have — it catches power interruptions, logging failures, and sessions that ended early. Data completeness asks whether, within the epochs that exist, the expected observations are present. A file can score 100 % on the first and 70 % on the second: the receiver ran the whole session but tracked half the satellites it should have, which usually means an obstructed sky or a failing antenna.
Multipath is the one to watch on short sessions
MP1 and MP2 are linear combinations that isolate code multipath from the geometry. Values are in metres and lower is better; the 0.5 m pass threshold is conservative for geodetic-grade equipment on a clear site and forgiving for anything mounted near a wall, a vehicle, or a roof edge. Multipath is also the indicator that degrades most predictably with a bad monument, which makes it the most useful one to check before a re-observation is still cheap.
Cycle slips scale with session length
A raw slip count means nothing without knowing how long the session ran, so PosFlow normalises it: fewer than 5 slips per 1000 epochs earns full marks, under 10 earns most of them, and the score falls off from there. Expressed the other way — observations per slip — 1000 or more passes and under 400 fails. High slip rates on a static session point at the antenna or the sky; on a kinematic one they are often just the environment.
What the score cannot tell you
It says nothing about whether the coordinates are right. A file can score 95 and still be logged with the wrong antenna height, the wrong antenna type, or a marker name that does not match the station log — none of which a quality metric can see. Check the header separately, every time.
Where to see this on your own data
Crux runs the QC engines and returns these indicators with the grade for each. A free account covers files up to 100 MB and reports are downloadable. The scoring code is the same for every mode; only the set of indicators an engine can report differs.
References and methodology
Technical definitions follow the sources below. Worked examples are illustrative; product-specific thresholds are identified as PosFlow settings. For corrections, contact PosFlow with the article URL and the relevant specification.