11.10. Skew Diagnostics

pycsamt.emtools.skew documents how far an impedance tensor departs from a simple 1-D or 2-D response. It gives two complementary views of skew:

• Bahr skewness, written \(\eta\), computed directly from the complex impedance tensor Z.

• Phase-tensor skew, written \(\beta\), computed through the phase-tensor table and reported in degrees.

Use skew diagnostics before 2-D preparation, dimensionality decisions, strike interpretation, frequency masking, and inversion setup. A low skew band supports a 1-D/2-D approximation. A high skew band warns that the tensor may contain 3-D structure, galvanic distortion, local noise, or component geometry that should not be forced into a simple 2-D workflow without more checks. Every example below uses pyCSAMT’s bundled AMT line, data/AMT/WILLY_DATA/L18PLT – a genuinely high-skew survey, as the numbers throughout this page will show.

Full signatures and parameter defaults are maintained in the API reference. This page explains how to use the public workflow functions through pycsamt.emtools.

11.10.1. Two Skew Measures

Bahr skewness is computed from impedance invariants. In pyCSAMT it is available through bahr_skewness and accepts either a tensor block of shape (n_freq, 2, 2) or a flattened block of shape (n_freq, 4). It is exactly the Bahr skewness glossary formula,

\[\eta = \sqrt{ \frac{|Z_{xx}+Z_{yy}|^2 + |Z_{xy}-Z_{yx}|^2} {|Z_{xx}-Z_{yy}|^2 + |Z_{xy}+Z_{yx}|^2} }\]

Phase-tensor skew starts from the phase tensor itself, built from the real and imaginary parts of the same impedance tensor,

\[\Phi = X^{-1} Y, \qquad X = \mathrm{Re}(Z), \quad Y = \mathrm{Im}(Z),\]

and the skew angle is then the Caldwell-Bibby-Bahr invariant, with \(\Phi = \begin{pmatrix}\Phi_{xx} & \Phi_{xy}\\ \Phi_{yx} & \Phi_{yy}\end{pmatrix}\),

\[\beta = \frac{1}{2}\operatorname{atan2}\!\bigl( \Phi_{xy} - \Phi_{yx},\ \Phi_{xx} + \Phi_{yy}\bigr),\]

reported in degrees – the same \(\beta\) the skew glossary entry defines, and rotation-invariant unlike the companion orientation angle \(\alpha = \tfrac{1}{2}\operatorname{atan2}(\Phi_{xy} + \Phi_{yx},\ \Phi_{xx} - \Phi_{yy})\) that build_phase_tensor_table also returns. skew_table returns the same table produced by the phase-tensor workflow, including station, freq, period, beta, and skew – in this module skew is just an alias for beta, kept because that is the column name the masking and voting helpers below were written against.

The two measures are related but not interchangeable. Bahr \(\eta\) is a direct invariant of Z. Phase-tensor \(\beta\) depends on the real and imaginary impedance parts through the phase tensor. They can agree that a station is non-2-D while ranking the severity differently.

11.10.2. Workflow Map

Goal	Use this	Result
Compute a station-frequency skew table	skew_table	A pandas.DataFrame with one row per station and frequency.
Compute Bahr skew from a Z block	bahr_skewness	A one-dimensional array of \(\eta\) values.
Mask rows above a skew threshold	mask_by_skew	A Sites object where rejected Z/tipper rows are NaN.
Keep one contiguous low-skew run per station	keep_longest_low_skew	A station-wise band mask.
Bridge short gaps inside low-skew runs	close_skew_gaps	A less fragmented station-wise mask.
Keep a survey-wide low-skew band	select_low_skew_band	A shared frequency band selected by station vote.
Plot skew quality across the line	plot_skew_traffic_psection, plot_skew_percentile_ribbon, plot_skew_vote_band	Pseudo-section, period ribbon, and vote-curve views.
Plot Bahr skew for one station	plot_skewness	A single-station \(\eta\) plot against a threshold.

11.10.3. Loading The Survey

Load once with ensure_sites and reuse the same raw object for diagnostics and before/after checks.

>>> from pathlib import Path
>>> from pycsamt.emtools import ensure_sites
>>> edi_dir = Path("data/AMT/WILLY_DATA/L18PLT")
>>> sites = ensure_sites(edi_dir, recursive=True, verbose=0)

Most masking functions default to inplace=False. That means the returned Sites object is masked, while sites remains available for comparison.

11.10.4. Phase-Tensor Skew Table

Start with skew_table. It summarizes \(\beta\) at every station-frequency row.

>>> from pycsamt.emtools import ensure_sites, skew_table
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> table = skew_table(sites)
>>> table.columns
Index(['station', 'freq', 'period', 's1', 's2', 'theta', 'alpha', 'beta',
       'skew', 'ellipt'],
      dtype='object')
>>> table[["station", "freq", "period", "beta", "skew"]].head()
   station     freq    period      beta      skew
0  18-001A  10400.0  0.000096  2.611588  2.611588
1  18-001A   8707.0  0.000115  1.964321  1.964321
2  18-001A   7289.0  0.000137  1.804266  1.804266
3  18-001A   6102.0  0.000164  1.068855  1.068855
4  18-001A   5108.0  0.000196 -6.949269 -6.949269
>>> table["skew"].abs().describe()
count    1484.000000
mean       23.850491
std        23.490731
min         0.050944
25%         5.375451
50%        14.920270
75%        37.654544
max        89.811834
Name: skew, dtype: float64
>>> station_summary = (
...     table.assign(abs_skew=table["skew"].abs())
...     .groupby("station", as_index=False)
...     .agg(
...         n_freq=("freq", "size"),
...         median_abs_skew=("abs_skew", "median"),
...         max_abs_skew=("abs_skew", "max"),
...     )
...     .sort_values("median_abs_skew")
... )
>>> station_summary.head()
    station  n_freq  median_abs_skew  max_abs_skew
0   18-001A      53         4.809977     88.625613
1   18-002U      53         6.596544     54.370326
6   18-007U      53         8.377503     54.171200
11  18-012A      53         9.052529     82.530092
7   18-008U      53         9.136982     81.792497
>>> station_summary.tail()
    station  n_freq  median_abs_skew  max_abs_skew
19  18-020A      53        45.539049     84.467238
21  18-021U      53        46.582007     79.328230
20  18-021B      53        47.235914     89.119132
22  18-022U      53        49.751057     89.263622
24  18-023A      53        54.529465     88.439189

Interpret skew and beta as angles in degrees. A common strict phase-tensor threshold is around 3 to 6 degrees. This line sits well above it: median absolute skew across all 1484 station-frequency rows is nearly 15 degrees, and the top quartile runs past 37. The two ends of station_summary also line up with where a station sits along the profile – the lowest-skew stations are all in the 18-001A-18-012A range, while every station from 18-020A onward lands in the highest-skew tail. When most of a survey clears a strict threshold by this much, do not blindly mask the whole line. First inspect the distribution and decide whether the threshold is appropriate for the question being asked.

11.10.5. Bahr Skewness

Use bahr_skewness when you want an independent skew measure computed directly from impedance.

>>> import numpy as np
>>> from pycsamt.emtools import bahr_skewness, ensure_sites
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> station = next(iter(sites))
>>> z = station.z
>>> freq = station.freq
>>> eta = bahr_skewness(z)
>>> period = 1.0 / freq
>>> print(np.nanmin(eta), np.nanmedian(eta), np.nanmax(eta))
0.8461749951030751 1.5856487464775848 2.9879890035270154
>>> print(period[:5], eta[:5])
[9.61538462e-05 1.14850121e-04 1.37193031e-04 1.63880695e-04
 1.95771339e-04] [2.97741225 2.987989   2.86896122 2.57559637 2.18481788]

The classic Bahr threshold often used as a 2-D/3-D guide is eta = 0.4. Every one of this station’s printed values sits well above it – even the minimum, 0.85, is more than double the threshold. Treat that boundary as a diagnostic aid, not an automatic editing rule. Large \(\eta\) can reflect genuine 3-D structure, but it can also reflect noise, distortion, or component problems.

11.10.6. Bahr Skew Plot

plot_skewness plots Bahr \(\eta\) against log-period and draws the threshold line.

>>> import matplotlib.pyplot as plt
>>> from pycsamt.emtools import ensure_sites, plot_skewness
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> station = next(iter(sites))
>>> fig, ax = plt.subplots(figsize=(7, 4))
>>> _ = plot_skewness(
...     station.freq,
...     station.z,
...     threshold=0.4,
...     ax=ax,
...     title=str(getattr(station, "name", "station")),
... )
>>> fig.tight_layout()
>>> fig.savefig("bahr_skewness_18-001A.png", dpi=200)
>>> plt.close(fig)

Every point for 18-001A sits above the red eta = 0.4 line, consistent with the printed minimum of 0.85 above – the annotated average, Bahr = 1.665, is itself more than four times the threshold. Use this plot when you need to explain one station concretely. Use the survey-wide phase-tensor plots below when the question is line-scale dimensionality.

11.10.7. Masking By Skew

mask_by_skew applies a phase-tensor skew threshold row by row. Every mode reduces to the same shape – keep the row unless \(\beta\) crosses a one-sided or two-sided bound – but which side of \(\beta\) gets rejected changes with the mode:

Mode	Keep rule	Use it to
"abs_gt" (default)	\(|\beta| \le \text{thresh}\)	reject rows too skewed in either direction – the usual symmetric threshold.
"gt"	\(\beta \le \text{thresh}\)	reject only large positive skew, keeping large negative skew.
"lt"	\(\beta \ge \text{thresh}\)	reject only large negative skew, keeping large positive skew.
"abs_lt"	\(|\beta| \ge \text{thresh}\)	the inverse of the default: keep only the strongly skewed rows, useful for isolating high-skew examples rather than removing them.

"gt" and "lt" are asymmetric on purpose. They let a survey tolerate skew of one sign – say, from a known distortion direction – while still rejecting the other.

>>> from pycsamt.emtools import ensure_sites, mask_by_skew
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> masked = mask_by_skew(
...     sites,
...     thresh=6.0,
...     mode="abs_gt",
...     also="both",
...     inplace=False,
... )
>>> n_before = sum(int(getattr(s, "z").shape[0]) for s in sites)
>>> import numpy as np
>>> n_after = sum(int(np.isfinite(getattr(s, "z")).all(axis=(1, 2)).sum()) for s in masked)
>>> print(n_before, "rows total,", n_after, "rows survive |beta| <= 6")
1484 rows total, 397 rows survive |beta| <= 6

Barely more than a quarter of this survey’s rows clear a strict 6-degree threshold. That is a legitimate outcome for a genuinely high-skew line – not a sign the function misbehaved – and it is exactly why the next several sections build gentler alternatives to a flat row-by-row cutoff.

also controls which data blocks are masked:

• "z" masks only impedance rows.

• "tipper" masks only tipper rows.

• "both" masks impedance and tipper rows at the same rejected frequencies.

The three non-default modes from the table above, in the same order:

>>> from pycsamt.emtools import mask_by_skew

>>> # Keep only rows with skew <= +6 degrees.
>>> keep_not_greater = mask_by_skew(sites, thresh=6.0, mode="gt")

>>> # Keep only rows with skew >= -6 degrees.
>>> keep_not_less = mask_by_skew(sites, thresh=-6.0, mode="lt")

>>> # Keep rows where absolute skew is at least 6 degrees.
>>> # This is unusual for inversion preparation, but useful for isolating
>>> # high-skew examples in diagnostics.
>>> keep_high_skew = mask_by_skew(sites, thresh=6.0, mode="abs_lt")

Because masking writes NaN into rejected tensor rows, always verify how many usable rows remain before sending the result to an inversion or plotting workflow.

11.10.8. Counting Surviving Rows

A simple count helper is often enough for documentation and QC tables.

>>> import numpy as np
>>> from pycsamt.emtools import ensure_sites, mask_by_skew
>>> raw = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> masked = mask_by_skew(raw, thresh=6.0, also="z", inplace=False)
>>> rows = []
>>> for station in masked:
...     z = station.z
...     good = np.isfinite(z).all(axis=(1, 2))
...     rows.append(
...         {
...             "station": getattr(station, "name", ""),
...             "n_total": int(z.shape[0]),
...             "n_kept": int(good.sum()),
...             "kept_fraction": float(good.mean()),
...         }
...     )
...
>>> rows[:5]
[{'station': '18-001A', 'n_total': 53, 'n_kept': 28, 'kept_fraction': 0.5283018867924528}, {'station': '18-002U', 'n_total': 53, 'n_kept': 25, 'kept_fraction': 0.4716981132075472}, {'station': '18-003A', 'n_total': 53, 'n_kept': 19, 'kept_fraction': 0.3584905660377358}, {'station': '18-004A', 'n_total': 53, 'n_kept': 16, 'kept_fraction': 0.3018867924528302}, {'station': '18-005U', 'n_total': 53, 'n_kept': 18, 'kept_fraction': 0.33962264150943394}]

Even 18-001A, the lowest-median-skew station in the whole survey, keeps barely over half its rows at this threshold. This count is the sanity check that prevents a silent all-NaN result from moving downstream. If strict thresholds keep too little data, return to the skew distribution and decide whether a looser survey-specific threshold is justified.

11.10.9. Longest Low-Skew Run

keep_longest_low_skew is stricter than row-by-row masking. For each station, it finds the longest contiguous frequency run satisfying abs(skew) <= thresh and masks everything else.

>>> from pycsamt.emtools import ensure_sites, keep_longest_low_skew
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> banded = keep_longest_low_skew(
...     sites,
...     thresh=6.0,
...     min_len=3,
...     pad=1,
...     also="both",
...     fallback="keep_all",
...     inplace=False,
... )
>>> import numpy as np
>>> kept = [int(np.isfinite(s.z).all(axis=(1, 2)).sum()) for s in banded]
>>> print("kept rows per station (first 8):", kept[:8])
kept rows per station (first 8): [15, 14, 16, 13, 13, 9, 9, 10]

min_len is important. If a station has only one or two isolated low-skew rows, you probably do not want to call that a usable band. pad expands the selected run by a few neighboring frequencies.

The fallback controls what happens when no acceptable run exists:

• fallback="keep_all" preserves the station rather than destroying it. This is safer for exploratory diagnostics, but it can make a failed threshold look like a fully accepted station unless you count rows.

• fallback="drop_all" masks the station completely when no run passes. This is stricter and should be used only when downstream code can handle missing stations or all-NaN rows.

11.10.10. Closing Small Gaps

close_skew_gaps fills short interior gaps inside a low-skew mask. It is useful when one or two noisy frequency samples break an otherwise continuous band.

>>> import numpy as np
>>> from pycsamt.emtools import close_skew_gaps, ensure_sites, mask_by_skew
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> before = mask_by_skew(sites, thresh=6.0, mode="abs_gt", also="z", inplace=False)
>>> gap_closed = close_skew_gaps(
...     sites,
...     thresh=6.0,
...     max_gap=1,
...     also="both",
...     inplace=False,
... )
>>> n_before = sum(int(np.isfinite(s.z).all(axis=(1, 2)).sum()) for s in before)
>>> n_after = sum(int(np.isfinite(s.z).all(axis=(1, 2)).sum()) for s in gap_closed)
>>> print("rows kept before gap-closing:", n_before)
rows kept before gap-closing: 397
>>> print("rows kept after gap-closing:", n_after)
rows kept after gap-closing: 422

Bridging single-sample gaps recovers 25 more rows here, spread across 15 of the 28 stations – real frequencies that a strict row-by-row cut had fragmented out of an otherwise continuous run. Set max_gap=0 to disable gap closing. Increase it cautiously. A large value can bridge unrelated good segments and create a band that no longer represents a genuinely continuous low-skew interval.

11.10.11. Survey-Wide Low-Skew Band

select_low_skew_band looks for frequencies supported by a fraction of stations. It first builds each station’s own accepted band exactly as keep_longest_low_skew would – the longest contiguous \(|\beta| \le \text{thresh}\) run, extended by pad – and only then projects every station’s band onto a shared union frequency grid \(G\) and votes:

\[\mathrm{keep}(f) = \frac{1}{N}\sum_{n=1}^{N} \mathbf{1}\bigl[f \in \mathrm{band}_n\bigr] \ \ge\ \text{frac}, \qquad f \in G,\]

where \(N\) is the number of stations and \(\mathrm{band}_n\) is station \(n\)’s own accepted run. A frequency survives only if it falls inside the contiguous accepted band of at least a frac-fraction of stations – not merely inside any low-skew row, which is the weaker condition the vote-band plot below checks.

>>> from pycsamt.emtools import ensure_sites, select_low_skew_band
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> shared = select_low_skew_band(
...     sites,
...     thresh=6.0,
...     frac=0.6,
...     min_len=3,
...     pad=0,
...     also="both",
...     inplace=False,
... )
>>> import numpy as np
>>> kept = [int(np.isfinite(s.z).all(axis=(1, 2)).sum()) for s in shared]
>>> print("total kept:", sum(kept), "of", sum(int(s.z.shape[0]) for s in sites))
total kept: 0 of 1484
>>> shared_loose = select_low_skew_band(
...     sites,
...     thresh=12.0,
...     frac=0.4,
...     min_len=3,
...     pad=0,
...     also="both",
...     inplace=False,
... )
>>> kept_loose = [int(np.isfinite(s.z).all(axis=(1, 2)).sum()) for s in shared_loose]
>>> print("loosened (thresh=12, frac=0.4) total kept:", sum(kept_loose))
loosened (thresh=12, frac=0.4) total kept: 364

At the function’s own defaults, the survey-wide vote finds nothing: the best-supported frequency has only 14 of 28 stations agreeing on their own contiguous low-skew run, short of the 16.8 stations frac=0.6 requires. That is a real, honest result for a line this skewed, not a bug – it is exactly the “no clean shared window” finding the traffic-light and ribbon plots below will confirm visually. Loosen the criteria (here, thresh=12 and frac=0.4) and a genuine shared band reappears, 364 rows this time. Use this function when you need one common frequency band for a line, for example before a shared inversion setup or a line-scale dimensionality statement; because it votes on contiguous per-station bands rather than isolated pointwise passes, it can legitimately return nothing even when individual stations and periods look locally fine.

11.10.12. Traffic-Light Pseudo-Section

plot_skew_traffic_psection colors each station-period cell by absolute phase-tensor skew:

• green: abs(beta) <= t1;

• amber: t1 < abs(beta) <= t2;

• red: abs(beta) > t2.

>>> from pycsamt.emtools import ensure_sites, plot_skew_traffic_psection
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> fig, ax = plt.subplots(figsize=(10, 5))
>>> _ = plot_skew_traffic_psection(
...     sites,
...     t1=3.0,
...     t2=6.0,
...     axis_y="logperiod",
...     ax=ax,
... )
>>> fig.tight_layout()
>>> fig.savefig("skew_traffic_psection_l18plt.png", dpi=200)
>>> plt.close(fig)

With strict t1=3/t2=6 thresholds, roughly 73 percent of this survey’s cells read red, 12 percent amber, and 14 percent green – visible here as a pseudo-section that is mostly pale red with only a scattering of stronger colour, rather than solid red throughout. The pattern is not random: the left half of the line (18-001A through roughly 18-015U) is noticeably paler than the right half (18-017U onward), the same early-versus-late split the station summary table showed earlier. For highly skewed surveys like this one, strict textbook thresholds are still useful – they tell you plainly that a clean 2-D band does not exist without relaxing the criteria or restricting to part of the line.

11.10.13. Percentile Ribbon

plot_skew_percentile_ribbon summarizes the whole line by period. It plots median absolute skew and percentile bands through the period axis.

>>> from pycsamt.emtools import ensure_sites, plot_skew_percentile_ribbon
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> fig, ax = plt.subplots(figsize=(8, 4))
>>> _ = plot_skew_percentile_ribbon(
...     sites,
...     n_bins=30,
...     q_lo=25.0,
...     q_hi=75.0,
...     extra=(10.0, 90.0),
...     ax=ax,
... )
>>> fig.tight_layout()
>>> fig.savefig("skew_percentile_ribbon_l18plt.png", dpi=200)
>>> plt.close(fig)

The median curve is genuinely period-dependent here: it stays under 15 degrees for periods shorter than about \(10^{-3}\,\mathrm{s}\), climbs into the 15-20 degree range through the middle of the band, and rises again toward 40-45 degrees at the longest periods – a real window rather than a uniformly high ribbon. Use this plot to answer: “Is there any period range where the line as a whole becomes less skewed?” Here, the answer is yes at short period, though even that window sits above a strict 3-6 degree threshold.

11.10.14. Vote-Band Plot

plot_skew_vote_band bins the survey in log-period, and within each bin \(b\) a station is counted as passing if a majority of its own rows in that bin are low-skew:

\[ \begin{align}\begin{aligned}\mathrm{pass}_n(b) = \frac{1}{|R_{n,b}|}\sum_{f \in R_{n,b}} \mathbf{1}\bigl[|\beta(f)| \le \text{thresh}\bigr] \; > \; 0.5,\\\qquad \mathrm{vote}(b) = \frac{1}{N}\sum_{n=1}^{N} \mathbf{1}\bigl[ \mathrm{pass}_n(b)\bigr],\end{aligned}\end{align} \]

where \(R_{n,b}\) is station \(n\)’s rows falling in bin \(b\). This is a pointwise vote – it never checks whether a station’s low-skew rows in neighboring bins are contiguous, which is exactly what select_low_skew_band requires above. A period range can therefore look well-supported here while still failing the stricter survey-wide band selection.

>>> from pycsamt.emtools import ensure_sites, plot_skew_vote_band
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> fig, ax = plt.subplots(figsize=(8, 4))
>>> _ = plot_skew_vote_band(
...     sites,
...     thresh=6.0,
...     n_bins=40,
...     ax=ax,
... )
>>> fig.tight_layout()
>>> fig.savefig("skew_vote_band_l18plt.png", dpi=200)
>>> plt.close(fig)

The vote fraction peaks near 0.9 at the very shortest periods, then falls fast and spends most of the band oscillating under 0.4 – never reaching the 0.6 support select_low_skew_band demanded by default above. That is consistent with the empty result found there: plenty of individual bins have some majority-passing stations, but not enough overlap between which stations pass at which periods to satisfy a stricter, contiguity-aware vote. Treat this plot as diagnostic rather than a substitute for select_low_skew_band: a high vote here says many stations are individually low-skew somewhere in the bin, not that they share one continuous band.

11.10.15. Suggested Interpretation Pattern

A robust skew review usually follows this sequence.

>>> from pycsamt.emtools import (
...     ensure_sites,
...     plot_skew_percentile_ribbon,
...     plot_skew_traffic_psection,
...     plot_skew_vote_band,
...     skew_table,
... )
>>> sites = ensure_sites("data/AMT/WILLY_DATA/L18PLT", recursive=True)
>>> table = skew_table(sites)
>>> table["skew"].abs().describe()
count    1484.000000
mean       23.850491
std        23.490731
min         0.050944
25%         5.375451
50%        14.920270
75%        37.654544
max        89.811834
Name: skew, dtype: float64
>>> fig, axes = plt.subplots(3, 1, figsize=(10, 12))
>>> _ = plot_skew_traffic_psection(sites, t1=3.0, t2=6.0, ax=axes[0])
>>> _ = plot_skew_percentile_ribbon(sites, ax=axes[1])
>>> _ = plot_skew_vote_band(sites, thresh=6.0, ax=axes[2])
>>> fig.tight_layout()
>>> fig.savefig("skew_review_panels_l18plt.png", dpi=200)
>>> plt.close(fig)

Read the three panels together, top to bottom: the pseudo-section shows where on the line and at which periods skew is worst (the right-hand, longer-period red block); the ribbon shows that the survey median genuinely improves at short period rather than staying flat; and the vote-band curve shows that even the best-supported period bins fall well short of the majority needed for a shared band. Only after reading all three together – not the summary table alone – should you choose mask_by_skew, keep_longest_low_skew, close_skew_gaps, or select_low_skew_band, and with which thresholds.

11.10.16. Pitfalls

High skew is not automatically bad data. It may be the geologic signal. Use skew masks as interpretation tools, not as a reflexive cleaning step.

Do not compare Bahr \(\eta\) and phase-tensor \(\beta\) as if they have the same units. \(\eta\) is dimensionless; \(\beta\) is an angle in degrees.

Beware of fallback behavior in keep_longest_low_skew. If fallback="keep_all", a station that fails the threshold completely can return with all rows preserved. Count the surviving rows and inspect the skew table.

Use also="z" when only impedance should drive downstream processing. Use also="both" when tipper rows must stay aligned with impedance masking.

11.10.17. Worked Example

The example uses the L18PLT survey to compare phase-tensor skew and Bahr skewness, apply skew-based masks, keep contiguous low-skew bands, close small gaps, vote on a shared low-skew band, and generate the traffic-light, percentile-ribbon, vote-band, and Bahr-skew figures.

Open the rendered gallery page here: Phase-tensor and Bahr skew diagnostics (pycsamt.emtools.skew).