11.3. Frequency Editing, Resampling, And QC#

pycsamt.emtools.frequency is the user-guide hub for operations that change, evaluate, or visualize the frequency axis. It is used before inversion, before section plotting, after quality control, and whenever multiple stations need to share a consistent frequency grid.

The module covers four practical jobs:

  • selecting a usable frequency band;

  • dropping, masking, or recovering low-confidence rows;

  • regridding, decimating, smoothing, and aligning station grids;

  • plotting frequency coverage, apparent depth, and band-level summary diagnostics.

Full callable signatures live in the API reference. This page focuses on how to use the workflows safely, how to read their tables, and how to avoid common frequency-axis traps.

11.3.1. Why Frequency Editing Matters#

CSAMT/AMT/MT processing is rarely limited by one bad station value. More often, a workflow succeeds or fails because the frequency axis is inconsistent, too noisy at the band edges, or incompatible across stations. A 2-D inversion, pseudo-section, or model-comparison workflow needs clear decisions about:

  • which band is physically meaningful;

  • which frequency rows are trusted;

  • whether bad rows are removed or kept as missing values;

  • whether stations should be interpolated to a shared grid;

  • whether smoothing or decimation is acceptable for the next step.

The frequency module makes those decisions explicit and auditable.

11.3.2. Data Contract#

All editing functions accept the usual emtools input:

  • a directory containing EDI files,

  • one EDI-like object,

  • a Sites container,

  • an iterable of site-like objects.

Internally the functions call ensure_sites. The familiar recursive, on_dup, strict, and verbose arguments behave the same way as in the rest of the user guide.

Most editing functions default to inplace=False. Keep that default while exploring. Switch to inplace=True only when you intentionally want to mutate the input object.

11.3.3. Selecting A Band#

Use select_band when you already know the frequency limits.

>>> from pathlib import Path
>>> from pycsamt.emtools.frequency import select_band
>>> edi_dir = Path("data/AMT/WILLY_DATA/L18PLT")
>>> mid_band = select_band(
...     edi_dir,
...     fmin=10.0,
...     fmax=1000.0,
...     inplace=False,
... )

fmin and fmax are in hertz. The convenience argument band_hz=(10.0, 1000.0) is also accepted. If both forms are provided, explicit fmin and fmax take precedence.

Use keep when you want specific frequencies rather than a continuous band.

>>> selected = select_band(
...     "data/AMT/WILLY_DATA/L18PLT",
...     keep=[10.0, 100.0, 1000.0],
... )

Band selection removes rows outside the chosen band. It changes the frequency grid length.

11.3.4. Removing Duplicate Frequencies#

Use drop_duplicates when a station has repeated or unsorted frequency rows.

>>> from pycsamt.emtools.frequency import drop_duplicates
>>> cleaned = drop_duplicates(
...     "data/AMT/WILLY_DATA/L18PLT",
...     tol=1e-10,
...     inplace=False,
... )

The function keeps the first occurrence of each frequency within the tolerance and applies the same row selection to impedance and tipper blocks when present.

11.3.5. Frequency Confidence#

Confidence-based editing uses frequency_confidence_table from the QC workflow (Quality-Control Confidence Scoring). Every station-frequency sample is scored with the same composite confidence ratio defined there, which folds six bounded diagnostics into one number:

\[\mathrm{CR}_{i,f} = { \sum_k w_k\, s_{k,i,f}\, \mathbf{1}_{s_{k,i,f}\ \mathrm{finite}} \over \sum_k w_k\, \mathbf{1}_{s_{k,i,f}\ \mathrm{finite}} }, \qquad 0 \le s_k \le 1,\]

where \(i\) indexes the station, \(f\) the frequency, and \(k\) runs over coverage, uncertainty, offdiag, diagonal, phase, and spatial, weighted respectively 0.35, 0.20, 0.15, 0.10, 0.10, 0.10. A missing component drops out of both the numerator and the denominator, so a station with no tipper is not penalized twice for the same gap. ci_hi and ci_lo then cut the [0, 1] range of \(\mathrm{CR}\) into three working classes: at or above ci_hi a row is trusted and kept as-is, between ci_lo and ci_hi it is a candidate for interpolation, and below ci_lo it is a rejection candidate. The frequency module computes this table internally, but you should look at its distribution before editing anything.

>>> import matplotlib.pyplot as plt
>>> from pycsamt.emtools.qc import frequency_confidence_table
>>> survey = "data/AMT/WILLY_DATA/L18PLT"
>>> confidence = frequency_confidence_table(survey, api=True).to_pandas()
>>> confidence[["station", "frequency_hz", "confidence", "flags"]].head()
  station  ...                                              flags
0  18-001A  ...  recoverable,high_error,offdiag_mismatch,diagon...
1  18-001A  ...  reject,high_error,offdiag_mismatch,diagonal_le...
2  18-001A  ...  reject,high_error,offdiag_mismatch,diagonal_le...
3  18-001A  ...  reject,high_error,offdiag_mismatch,diagonal_le...
4  18-001A  ...  reject,high_error,offdiag_mismatch,diagonal_le...
[5 rows x 4 columns]
>>> confidence["confidence"].describe()
count    1484.000000
mean        0.659129
std         0.096056
min         0.417464
25%         0.587712
50%         0.656305
75%         0.737060
max         0.863041
Name: confidence, dtype: float64
>>> station = "18-001A"
>>> one = confidence.loc[confidence["station"] == station].sort_values("period_s")
>>> fig, ax = plt.subplots(figsize=(7, 4.5))
>>> _ = ax.semilogx(one["period_s"], one["confidence"], "o-")
>>> _ = ax.axhline(0.90, color="tab:red", linestyle="--", label="ci_hi")
>>> _ = ax.axhline(0.50, color="tab:orange", linestyle="--", label="ci_lo")
>>> _ = ax.set_xlabel("Period (s)")
>>> _ = ax.set_ylabel("Confidence")
>>> _ = ax.legend()
>>> fig.tight_layout()
../../_images/user-guide-emtools-frequency-04.png

The key lesson is simple: do not choose ci_hi, ci_lo, or threshold before looking at the confidence distribution above. In this survey the median \(\mathrm{CR}\) sits near 0.66 and the maximum barely reaches 0.86. Leaving ci_hi at a textbook value like 0.90 would mean no row is ever trusted enough to act as an interpolation donor, and recovery would silently have nothing to work from. Read the describe() output first, then set ci_hi and ci_lo (or threshold) to values this particular survey can actually reach.

11.3.6. Drop Or Mask Low-Confidence Rows#

Once a threshold is chosen, both functions apply the same keep rule to row \((i, f)\):

\[\text{keep}_{i,f} = \bigl(\mathrm{CR}_{i,f}\ \text{undefined}\bigr) \ \lor\ \bigl(\mathrm{CR}_{i,f} \ge \text{threshold}\bigr).\]

Rows with no confidence score at all (for example, a frequency present in the impedance block but absent from the confidence table) are kept rather than silently discarded — undefined is not the same as bad. What differs between the two functions is what happens to the rows that fail the rule. Dropping removes them from the frequency grid entirely. Masking keeps the grid the same length but replaces the corresponding tensor rows with NaN.

>>> from pycsamt.emtools.frequency import (
...     drop_low_confidence_frequencies,
...     mask_low_confidence_frequencies,
... )
>>> dropped = drop_low_confidence_frequencies(
...     survey,
...     threshold=0.50,
...     also="both",
...     inplace=False,
... )
>>> masked = mask_low_confidence_frequencies(
...     survey,
...     threshold=0.50,
...     also="both",
...     inplace=False,
... )

Use drop when downstream code cannot handle missing rows. Use mask when preserving the original frequency grid matters, for example when comparing before and after processing. Dropping shrinks the impedance array before the frequency array, in that order, to avoid a hard failure in the underlying setter; you may see one harmless ERROR-level log line per edited station as a result (“Failed to compute rho/phi after setting Z”) – it self-corrects one line later once the frequency array is updated to match, and does not indicate that anything actually failed.

The also argument controls which blocks are edited:

Value

Meaning

"z"

Edit impedance only.

"tipper"

Edit tipper only.

"both"

Edit impedance and tipper when present.

11.3.7. Recover Low-Confidence Rows#

recover_low_confidence_frequencies splits a station’s rows into three sets from the same \(\mathrm{CR}_{i,f}\) used above:

\[G = \{f : \mathrm{CR}_{i,f} \ge \text{ci\_hi}\}, \qquad R = \{f : \text{ci\_lo} \le \mathrm{CR}_{i,f} < \text{ci\_hi}\}, \qquad B = \{f : \mathrm{CR}_{i,f} < \text{ci\_lo}\},\]

the trusted donors \(G\), the recoverable rows \(R\), and the rejected rows \(B\). Every row in \(R\) is filled by interpolating the complex tensor value at that frequency from the rows in \(G\), in log-frequency space and independently for the real and imaginary parts:

\[Z(f_i) = \mathrm{interp}\bigl(\log_{10} f_i;\ \{\log_{10} f_g\}_{f_g \in G},\ \{Z(f_g)\}_{f_g \in G}\bigr), \qquad f_i \in R,\]

with interpolation="linear" performing ordinary linear interpolation in \(\log_{10} f\) and "nearest" copying the closest donor instead. Working in \(\log_{10} f\) rather than raw hertz matters because CSAMT/AMT bands are sampled logarithmically — linear interpolation in hertz would be dominated by the widely spaced high-frequency end and would barely constrain the low-frequency end at all. Rows in \(B\) are not interpolated; they are handled entirely by reject.

>>> from pycsamt.emtools.frequency import recover_low_confidence_frequencies
>>> recovered = recover_low_confidence_frequencies(
...     "data/AMT/WILLY_DATA/L18PLT",
...     ci_hi=0.72,
...     ci_lo=0.50,
...     interpolation="linear",
...     reject="drop",
...     also="both",
...     inplace=False,
... )

Valid interpolation modes are "linear" and "nearest". Valid reject modes are:

Reject mode

Meaning

"drop"

Remove rows below ci_lo.

"mask"

Keep rows below ci_lo but set them to NaN.

"keep"

Leave rejected rows unchanged.

Recovery is interpolation, not magic: it can only be as good as the donors in \(G\). When a recoverable row sits outside the span of \(\log_{10} f_g\) covered by the donors, interp clamps to the nearest edge, so “linear” quietly degrades to “nearest” exactly at the band edges where recovery is riskiest. Always inspect the result when many rows are recovered, and treat edge-of-band recoveries with more suspicion than interior ones.

11.3.8. High-Level Editing Workflow#

Use edit_frequencies_by_confidence when you want the edited sites, a station-level report, and a row-level decision table in one object.

>>> from pycsamt.emtools.frequency import edit_frequencies_by_confidence
>>> result = edit_frequencies_by_confidence(
...     "data/AMT/WILLY_DATA/L18PLT",
...     mode="recover",
...     ci_hi=0.72,
...     ci_lo=0.50,
...     reject="drop",
...     interpolation="linear",
...     api=False,
... )
>>> result.summary()
"FrequencyEditResult(mode='recover', method='composite', dropped=68, masked=102, recovered=867)"
>>> result.report.head()
   station  n_freq_before  ...  n_masked_or_unfinite  confidence_delta
0  18-001A             53  ...                     0          0.027413
1  18-002U             53  ...                     0         -0.038885
2  18-003A             53  ...                     3          0.061644
3  18-004A             53  ...                     3         -0.026442
4  18-005U             53  ...                     1          0.026781
[5 rows x 18 columns]
>>> result.decisions.head()
   station  frequency_hz  period_s  ...  present_after  finite_after action
0  18-001A       10400.0  0.000096  ...           True          True   kept
1  18-001A        8707.0  0.000115  ...           True          True   kept
2  18-001A        7289.0  0.000137  ...           True          True   kept
3  18-001A        6102.0  0.000164  ...           True          True   kept
4  18-001A        5108.0  0.000196  ...           True          True   kept
[5 rows x 10 columns]

result is a FrequencyEditResult with:

  • sites: edited sites.

  • report: station-level before/after table.

  • decisions: one row per original station-frequency row.

  • n_dropped, n_masked, n_recovered: convenience counts.

  • mode, method, ci_hi, ci_lo, reject, interpolation: edit metadata.

For in-memory objects, pass an independent before_sites when you need an untouched baseline for reporting. Some lower-level editors may mutate wrapped impedance objects while constructing edited outputs.

11.3.9. Station-Level Report#

frequency_edit_report compares before and after sites.

>>> from pycsamt.emtools.frequency import (
...     edit_frequencies_by_confidence,
...     frequency_edit_report,
... )
>>> survey = "data/AMT/WILLY_DATA/L18PLT"
>>> result = edit_frequencies_by_confidence(
...     survey,
...     mode="drop",
...     threshold=0.50,
...     api=False,
... )
>>> report = frequency_edit_report(
...     survey,
...     result.sites,
...     ci_hi=0.90,
...     ci_lo=0.50,
...     api=True,
... ).to_pandas()
>>> report[[
...     "station", "n_freq_before", "n_freq_after", "n_dropped",
...     "n_finite_before", "n_finite_after",
...     "confidence_median_before", "confidence_median_after",
... ]].head()
   station  n_freq_before  ...  confidence_median_before  confidence_median_after
0  18-001A             53  ...                  0.711753                 0.711753
1  18-002U             53  ...                  0.749480                 0.749480
2  18-003A             53  ...                  0.666613                 0.671639
3  18-004A             53  ...                  0.735994                 0.743510
4  18-005U             53  ...                  0.728841                 0.733250
[5 rows x 8 columns]

Important report columns include:

  • n_freq_before and n_freq_after: grid size change.

  • n_finite_before and n_finite_after: finite tensor row counts.

  • frac_finite_before and frac_finite_after: finite row fraction.

  • safe_fraction_*: fraction at or above ci_hi.

  • recoverable_fraction_*: fraction in [ci_lo, ci_hi).

  • reject_fraction_*: fraction below ci_lo.

  • n_dropped: number of removed frequency rows.

  • n_masked_or_unfinite: finite rows lost to masking or missing data.

  • confidence_delta: median confidence change.

11.3.10. Decision Table#

frequency_edit_decision_table records one row per original station-frequency sample.

>>> from pycsamt.emtools.frequency import (
...     edit_frequencies_by_confidence,
...     frequency_edit_decision_table,
... )
>>> survey = "data/AMT/WILLY_DATA/L18PLT"
>>> result = edit_frequencies_by_confidence(
...     survey,
...     mode="recover",
...     ci_hi=0.72,
...     ci_lo=0.50,
...     reject="drop",
...     api=False,
... )
>>> decisions = frequency_edit_decision_table(
...     survey,
...     result.sites,
...     ci_hi=0.72,
...     ci_lo=0.50,
...     api=True,
... ).to_pandas()
>>> decisions["action"].value_counts()
action
recovered    867
kept         447
masked       102
dropped       68
Name: count, dtype: int64

Actions are:

  • kept: row survived unchanged.

  • dropped: row is absent after editing.

  • masked: row remains but is no longer finite.

  • recovered: row was filled or changed by recovery.

11.3.11. Plot Edit Results#

Use the summary plot for station-level effects and the decision plot for row-level actions.

>>> from pycsamt.emtools.frequency import (
...     plot_frequency_edit_decisions,
...     plot_frequency_edit_summary,
... )
>>> fig, (ax_summary, ax_decisions) = plt.subplots(2, 1, figsize=(10, 8))
>>> _ = plot_frequency_edit_summary(
...     survey, result.sites, ci_hi=0.72, ci_lo=0.50, ax=ax_summary,
... )
>>> _ = plot_frequency_edit_decisions(
...     survey, result.sites, ci_hi=0.72, ci_lo=0.50, ax=ax_decisions,
... )
>>> fig.tight_layout()
../../_images/user-guide-emtools-frequency-10.png

If the decision plot is dominated by recovered or masked colors, the edit is not conservative and should be justified.

11.3.12. Regrid To A Target Grid#

Use regrid_to when you already have the target frequencies.

>>> import numpy as np
>>> from pycsamt.emtools.frequency import regrid_to
>>> target_freq = np.array([1.0, 3.0, 10.0, 30.0, 100.0, 300.0, 1000.0])
>>> regridded = regrid_to(
...     "data/AMT/WILLY_DATA/L18PLT",
...     target_freq,
...     method="linear",
...     inplace=False,
... )

regrid_to reuses the same log-frequency interpolator introduced above for recovery, except every target frequency is filled from the entire native grid rather than from a confidence-selected subset of it — there is no trusted/recoverable split here, only the target grid you hand it. method="linear" interpolates the real and imaginary parts in \(\log_{10} f\); method="nearest" copies the closest native row instead of interpolating. Choose a target grid that is meaningful for the inversion or comparison you are preparing for, not merely one that is convenient to type.

11.3.13. Build A Log-Spaced Grid#

Typing out a target grid by hand gets tedious once a survey needs a regular per-decade sampling. regrid_logspace builds one automatically by geometric progression:

\[f_k = f_{\min} \left(\frac{f_{\max}}{f_{\min}}\right)^{k / (n - 1)}, \qquad k = 0, \dots, n - 1, \qquad n = \max\!\left(2,\ \left\lceil \log_{10}(f_{\max} / f_{\min}) \times \text{per\_decade} \right\rceil\right).\]

Equal steps in \(k\) are equal steps in \(\log_{10} f\), so per_decade directly controls resolution: per_decade=6 places six points in every factor-of-ten span, matching the density a typical CSAMT/AMT band is actually sampled at. When fmin/fmax are omitted, they default to the min/max of the union grid across all stations.

>>> from pycsamt.emtools.frequency import regrid_logspace
>>> regular = regrid_logspace(
...     "data/AMT/WILLY_DATA/L18PLT",
...     fmin=1.0,
...     fmax=10000.0,
...     per_decade=6,
...     method="linear",
...     inplace=False,
... )

band_hz=(lo, hi) is accepted as an alias for fmin and fmax. n_per_decade is accepted as an alias for per_decade when per_decade is left at its default.

11.3.14. Decimation And Moving Average Smoothing#

decimate_step keeps rows at indices \(0, \text{step}, 2\cdot \text{step}, \dots\) along the native (not log-spaced) frequency array — it thins the grid but does not touch the tensor values that survive. smooth_mavg does the opposite: it keeps every row but replaces each tensor value with a centered moving average of width \(k\), computed independently on the real and imaginary parts,

\[Z'(f_i) = \frac{1}{k} \sum_{j=-\lfloor k/2 \rfloor}^{\lceil k/2 \rceil - 1} Z(f_{i+j}),\]

so it is a smoothing filter over row order, not over \(\log_{10} f\), and it changes every value it touches rather than removing any. Near the two ends of the grid the window runs past the available rows; the implementation still divides by the full \(k\) rather than shrinking the window, which pulls the first and last \(\lfloor k/2\rfloor\) smoothed values slightly toward zero. Keep k small relative to the number of rows in a station’s band if the edge rows still need to be trusted afterward.

>>> from pycsamt.emtools.frequency import decimate_step, smooth_mavg
>>> decimated = decimate_step(
...     "data/AMT/WILLY_DATA/L18PLT",
...     step=2,
...     inplace=False,
... )
>>> smoothed = smooth_mavg(
...     "data/AMT/WILLY_DATA/L18PLT",
...     k=3,
...     on="z",
...     inplace=False,
... )

Smoothing changes the measured tensor values. Use it as a processing choice, not as a plotting convenience, and report the window size.

11.3.15. Align Station Grids#

Use align_grid when stations need a common grid.

>>> from pycsamt.emtools.frequency import align_grid
>>> union_aligned = align_grid(
...     "data/AMT/WILLY_DATA/L18PLT",
...     mode="union",
...     method="nearest",
... )
>>> intersection_aligned = align_grid(
...     "data/AMT/WILLY_DATA/L18PLT",
...     mode="intersection",
...     method="nearest",
... )
>>> ref_aligned = align_grid(
...     "data/AMT/WILLY_DATA/L18PLT",
...     mode="ref",
...     ref_station="18-001A",
...     method="nearest",
... )

The modes are:

Mode

Meaning

union

Use all frequencies present in any station.

intersection

Use only frequencies present in every station.

ref

Use the grid from ref_station.

intersection can be empty or nearly empty for independently processed data because frequencies may not match bit-for-bit. If the intersection is empty, the function returns the input sites unchanged.

11.3.16. Coverage And Quality Heatmap#

plot_coverage_quality_heatmap shows which station-frequency cells exist and how reliable they are according to relative impedance error. This is a separate, lighter-weight score than the composite \(\mathrm{CR}\) used above — it looks only at the off-diagonal impedance error, so it stays cheap enough to compute for a full survey heatmap. For each cell it averages the relative error of the two off-diagonal modes and maps that to a quality in (0, 1]:

\[r_{i,f} = \frac{1}{2}\left( \frac{|\sigma_{Z_{xy}}|}{|Z_{xy}|} + \frac{|\sigma_{Z_{yx}}|}{|Z_{yx}|} \right), \qquad q_{i,f} = \frac{1}{1 + r_{i,f}},\]

where \(\sigma_Z\) is the reported impedance error at that cell. A noise-free measurement has \(r = 0\) and \(q = 1\); as the error grows relative to the signal, \(q\) decays smoothly toward 0 instead of blowing up, which keeps the colour scale usable even for badly noisy cells.

>>> from pycsamt.emtools.frequency import plot_coverage_quality_heatmap
>>> fig, ax = plt.subplots(figsize=(9, 4.5))
>>> _ = plot_coverage_quality_heatmap(
...     "data/AMT/WILLY_DATA/L18PLT",
...     axis="period",
...     ax=ax,
... )
>>> fig.tight_layout()
../../_images/user-guide-emtools-frequency-15.png

Values closer to 1 are higher quality; missing cells simply mean that station has no row at that union-grid frequency, which is a coverage gap rather than a quality problem.

11.3.17. Apparent Depth Pseudo-Section#

plot_apparent_depth_psection turns each station-frequency cell into a rough sense of investigation depth. It first needs an apparent resistivity, which it gets from the determinant of the two off-diagonal impedance tensor modes — the same geometric-mean convention used elsewhere in emtools (see CSUMT Bostick Depth And Survey Design):

\[\rho_{a,\det} = \sqrt{ \left(0.2\,\frac{|Z_{xy}|^2}{f}\right) \left(0.2\,\frac{|Z_{yx}|^2}{f}\right) },\]

and then converts that resistivity and frequency into a skin depth-style apparent depth,

\[\delta \approx 503 \sqrt{\rho_{a,\det} / f},\]

with \(\delta\) in metres when \(\rho_{a,\det}\) is in \(\Omega\,\mathrm{m}\) and \(f\) is in hertz. Lower frequencies and higher apparent resistivity both push \(\delta\) deeper, which is exactly the trade-off a CSAMT/AMT sounding is designed to exploit.

>>> from pycsamt.emtools.frequency import plot_apparent_depth_psection
>>> fig, ax = plt.subplots(figsize=(9, 4.5))
>>> _ = plot_apparent_depth_psection(
...     "data/AMT/WILLY_DATA/L18PLT",
...     axis_y="period",
...     agg="median",
...     log_color=True,
...     ax=ax,
... )
>>> fig.tight_layout()
../../_images/user-guide-emtools-frequency-16.png

This is a diagnostic visualization, not an inversion result — \(\delta\) is a proxy, not a layer boundary. Use it to see how apparent depth varies across stations and periods before committing to a model parameterization.

11.3.18. Band Microstrips#

A full pseudo-section is the right tool for inspecting one survey closely, but it stops being readable once you want to compare many lines side by side. plot_band_microstrips collapses the period axis into a handful of bands and reduces each station-band cell to three median summary metrics:

\[\log_{10}\rho_{a,\det}, \qquad \varphi_{\det} = \arg\bigl[\det(Z)\bigr] = \arg\bigl(Z_{xx}Z_{yy} - Z_{xy}Z_{yx}\bigr), \qquad |T| = \sqrt{|T_x|^2 + |T_y|^2},\]

the log-apparent-resistivity defined above, the determinant phase (in degrees, using the same \(\det(Z)\) convention as the tensor pages), and the tipper magnitude when a station has one. Each metric gets its own marker — circle, square, triangle — and its own colour normalization, so a single compact grid of dots can stand in for what would otherwise be three separate pseudo-sections per line.

>>> from pycsamt.emtools.frequency import plot_band_microstrips
>>> bands = [
...     (1e-4, 1e-3),
...     (1e-3, 1e-2),
...     (1e-2, 1e-1),
...     (1e-1, 1.0),
... ]
>>> fig, ax = plt.subplots(figsize=(9, 6))
>>> _ = plot_band_microstrips(
...     "data/AMT/WILLY_DATA/L18PLT",
...     bands=bands,
...     ax=ax,
... )
>>> fig.tight_layout()
../../_images/user-guide-emtools-frequency-17.png

Use microstrips for compact survey summaries. They are not a substitute for full curves when a band contains strong internal variation.

11.3.20. Common Failure Modes#

Using default ci_hi without checking confidence

Recovery needs trusted rows at or above ci_hi. If the survey confidence never reaches that value, recovery cannot behave as expected.

Dropping rows before checking downstream needs

Dropping shortens the grid. If the next step expects aligned grids, masking may be safer.

Masking rows before plotting finite-only workflows

Some workflows ignore or fail on NaN rows. Save a decision table so missing values are explainable.

Regridding too early

Interpolation can hide native acquisition problems. Inspect native confidence first.

Empty intersection alignment

Independently processed stations may have no exact shared frequency. Use union or a reference station when appropriate.

Smoothing as hidden preprocessing

Moving averages change tensor values. Report the window and apply it intentionally.

11.3.21. Worked Example#

The gallery example uses L18PLT from data/AMT/WILLY_DATA/ and brings in KAP03 for heterogeneous-grid alignment. It demonstrates band selection, confidence editing, recovery, reports, decision plots, regridding, decimation, smoothing, grid alignment, coverage heatmaps, apparent-depth pseudo-sections, and band microstrips.

Open the rendered example here: Frequency editing, resampling, and QC (pycsamt.emtools.frequency).