Routines

Routines are standalone, array-in / array-out processing functions. Each accepts array-like inputs – a scalar, a 1-D sequence (list or tuple), or a NumPy array, depending on what the routine needs – plus optional scalar parameters, and returns NumPy arrays; inputs are converted with numpy.asarray internally, so you need not build NumPy arrays yourself. Each routine documents (and validates) the exact input shape it expects. Each routine lives in its own versioned file, so it can be lifted into another project as a standalone module. A routine relies on the standard scientific stack (NumPy, SciPy, TEOS-10/GSW) where possible. If a routine needs a more specialised package, that package does not ship with pyRSKtools, so you can install it separately.

Note

Calling a routine. Import it from pyrsktools.routines:

from pyrsktools.routines import correcttempADCcal

temperature_corrected, mask = correcttempADCcal(temperature, time)

correcttempADCcal

pyrsktools.routines.correcttempADCcal(temperature: Vector, time: Vector) → Tuple[NDArray, NDArray]

Remove the periodic A2D self-calibration outliers from a temperature vector.

Version: 1.0.0

Parameters:

  • temperature – 1-D vector of temperature samples in degrees C (a list, tuple, or 1-D numpy array).

  • time – matching 1-D vector of sample times in seconds (same length as temperature). Elapsed or absolute both work: the routine phase-locks on the 60-s self-cal cycle, so it only needs the 60-s periodicity and a consistent phase – not an absolute wall-clock reference. The intervals must be the true sample spacing.

Returns:

tuple – (temperature_corrected, mask) – both are 1-D numpy.ndarray the same length as the inputs. temperature_corrected (float) is the cleaned series (still a valid temperature, a drop-in replacement) and mask (bool) is True for the samples that were actually corrected. A detected run that was skipped (no finite neighbour, or the whole series) or whose interpolated proposal clamped to no change is excluded, so mask.sum() is the number of samples genuinely modified. If the periodic self-cal signature cannot be confirmed (the file does not exhibit it, or is too short to verify), the series is returned unchanged with an all-False mask – the routine never fabricates a correction.

Raises:

ValueError – if the inputs are not matched 1-D vectors (a scalar or a 2-D/N-D input, or two different lengths), or if time decreases.

Example:

>>> from pyrsktools import routines
>>> temperature_corrected, mask = routines.correcttempADCcal(temperature, time)

inairO2calibration

pyrsktools.routines.inairO2calibration(dissolved_o2_concentration: Vector, odotemperature: Vector, patm: Vector, humidity: Vector) → Tuple[float, NDArray, NDArray, NDArray, NDArray]

In-air calibration gain for an optical dissolved-oxygen (ODO) sensor.

Version: 1.0.0

Computes the multiplicative gain that brings an optode’s in-air readings onto the atmospheric reference, following the SCOR WG 142 / Bittig in-air procedure. For each in-air sample the reference partial pressure po2_air is computed from atmospheric physics and compared with the optode’s partial pressure pO2; the per-sample ratio is m = po2_air / pO2 and the returned gain is its nan-mean.

This is a pure, standalone routine (no RSK object, no file I/O). The RSK-aware wrapper pyrsktools.RSK.inairO2calibration() prepares the inputs (in-air window selection and salinity un-compensation) and writes the results back to the file.

Caution

dissolved_o2_concentration MUST be salinity-uncompensated (S = 0). RBR ODO firmware salinity-compensates its output; that compensation must be stripped first, or the gain is biased. The routine cannot detect compensated input, so it emits a UserWarning on every call.

Parameters:

  • dissolved_o2_concentration – dissolved O2 concentration (µmol/L), salinity-uncompensated (S = 0); a 1-D vector (list, tuple, or numpy array).

  • odotemperature – the optode’s own temperature (°C); a 1-D vector the same length as dissolved_o2_concentration.

  • patm – atmospheric (sea-surface) pressure (dbar); a scalar or a matching 1-D vector.

  • humidity – relative humidity (%); a scalar or a matching 1-D vector.

Returns:

tuple – (gain, dissolved_o2_concentration_corrected, dissolved_o2_saturation_corrected, gain_series, po2_air)

  • gain (float) – scalar calibration, nanmean(po2_air / pO2_optode).

  • dissolved_o2_concentration_corrected (ndarray, µmol/L) – gain x concentration.

  • dissolved_o2_saturation_corrected (ndarray, %) – gain x optode saturation.

  • gain_series (ndarray) – per-sample gain m = po2_air / pO2 (a stability check); non-finite ratios (e.g. from a sample reading 0) are set to NaN and excluded from gain.

  • po2_air (ndarray, mbar) – reference atmospheric pO2.

The four arrays have the same length as dissolved_o2_concentration.

Raises:

ValueError – if odotemperature is not the same length as dissolved_o2_concentration, or if patm / humidity are neither scalar nor that length.

Example:

>>> from pyrsktools import routines
>>> gain, conc_corrected, sat_corrected, m, po2_air = routines.inairO2calibration(
...     dissolved_o2_concentration, odotemperature, patm=10.1325, humidity=80.0
... )

References

  • Bittig, H. C., et al. (2018), SCOR WG 142 in-air calibration procedure.

  • Garcia, H. E., & Gordon, L. I. (1992), Limnol. Oceanogr. 37(6), 1307-1312.

  • Weiss, R. F., & Price, B. A. (1980), Mar. Chem. 8, 347-359.