qa4sm-preprocessing 0.2

Preprocessing routines for QA4SM

This package contains functions to preprocess certain data before using them in the QA4SM online validation framework.

ISMN FRM

This module contains the routine to assign FRM qualifications ISMN sensors. The Quality Indicators (QIs) are based on a Triple Collocation run with 80% CI between ISMN (0-10 cm, “good” time stamps), ERA5-Land (“swvl1”) and ESA CCI SM v06.1 PASSIVE. See ./docs/ismn_frm.rst for more details.

CGLS HR SSM SWI

Read CGLS HR SSM and SWI images (1km sampling) in netcdf format and convert them to time series. The image reader allows reading/converting data for a spatial subset (bbox) only. Time series are stored in 5*5 DEG cell files, i.e. there are ~250 000 time series stored in one single cell file.

Time series reading is done based on cell level. Up to 6 cells are loaded into memory at a time. The read_area function allows reading multiple GPI time series around one location at once (and optionally converting them into a single, averaged series, to represent the mean SM for an area).

Necessary updates

At the moment it is only possible to read a single variable. However, in order to mask SM time series based in location quality flags, it is necessary to read multiple parameters. When passing the averaged time series for an area to pytesmo for validation, masking can not be done in pytesmo, but must be done beforehand.

Note

This project has been set up using PyScaffold 4.0.2. For details and usage information on PyScaffold see https://pyscaffold.org/.