Timeserie¶

pikobs.timeserie Module¶

The timeserie module within the pikobs package is designed for advanced time-series analysis of meteorological observations. It enables users to visualize and analyze both the quantity and quality of observations across customizable time periods, making it suitable for robust scientific and operational studies.

Main Features¶

  • Comprehensive Visualization: Generate time-series plots for observation counts, bias, and standard deviation.

  • Flexible Grouping Options: Analyze data per station, aggregated across all stations, or compare different configurations and satellites.

Example Visualizations¶

  • Disaggregated station statistics (id_stn=all):

    _images/ts.png

    Displays individual station statistics for more granular comparison.

Generated Graphs¶

  1. Bias and Standard Deviation - Bias: Indicates persistent differences between observations and reference data, highlighting systematic errors. - Standard Deviation: Represents the variability of observations over time.

  2. Observation Count Over Time - Illustrates the density and availability of observations, helping detect temporal gaps or inconsistencies.

Usage & Integration¶

This module can be executed from the command line, allowing seamless integration into batch workflows. Below is an example Bash script:

#!/bin/bash

# Locate Python executable
PYTHON_EXEC=$(which python)

# Run the timeserie analysis
$PYTHON_EXEC -c 'import pikobs; pikobs.timeserie.arg_call()' \

  --path_experience_files "/path/to/experience_files/"
  --experience_name "ops+NOAA21+GOES19"
  --path_control_files "/path/to/control_files/"
  --control_name "ops"
  --pathwork "onthefly2"
  --datestart "2025041200"
  --dateend "2025041300"
  --region "Monde"
  --family "sw"
  --flags_criteria "assimilee"
  --fonction "omp"
  --id_stn "join"
  --channel "join"
  --n_cpu 40

Parameter Descriptions¶

  • path_experience_files: Directory containing experience data files.

  • experience_name: Identifier for the specific experiment.

  • path_control_files: Directory containing control/reference data.

  • control_name: Identifier for the control/reference configuration.

  • pathwork: Working directory where outputs and intermediates are stored.

  • datestart: Start timestamp for the analysis period (format: YYYYMMDDHH).

  • dateend: End timestamp for the analysis period (format: YYYYMMDDHH).

  • region: Geographic area of interest. Valid values include: - Monde, PoleNord, PoleSud, AmeriqueduNord, OuestAmeriqueduNord, AmeriqueDuNordPlus, ExtratropiquesNord, HemisphereNord, HemisphereSud, Asie, Europe, Mexique, Canada, BaieDhudson, Arctiquecanadien, EtatsUnis, Tropiques30, Tropiques, Australie, Pacifique, Atlantique.

  • family: Observation family (e.g., mwhs2, to_amsua, iasi_qc, atms_qc, csr_qc, etc.).

  • flags_criteria: Filtering criterion (e.g., “assimilee”).

  • fonction: Computational function used (e.g., “omp”).

  • id_stn: Specifies stations or satellites to include (e.g., “all”, “join”, or specific satellite IDs like METOP-1).

  • channel: Specifies channels to analyze (“all”, “join”, or specific channel numbers).

  • n_cpu: Number of CPU cores to use for parallel processing.

Tip: Adjust the command-line arguments to match your data, region, and analysis goals.