Skip to content

Releases: NCAR/GARD

Version 1.0

26 May 14:20
@gutmann gutmann
v1.0
a095865
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
Version 1.0 Latest
Latest

This release represents the first complete version of the code as used in the published paper.

Gutmann, E. D., J. J. Hamman, M. P. Clark, T. Eidhammer, A. W. Wood, and J. R. Arnold, 2022: En-GARD: A Statistical Downscaling Framework to Produce and Test Large Ensembles of Climate Projections. J. Hydrometeor., 23, 1545–1561, https://doi.org/10.1175/JHM-D-21-0142.1.

What's Changed

  • Many small bugfixes
  • Addition of stochastic perturbation to analog fields to prevent identical values from always selecting the first dates
  • Improved calendar handling
  • Speed improvements to a few components
  • Internal code cleanup
  • adding ERA5 conversion file by @sam-hartke-ucar in #79
  • Update writeERA5file.py to run in parallel by @sam-hartke-ucar in #81

New Contributors

Full Changelog: v0.4...v1.0

Contributors

wcurrier and sam-hartke-ucar
Assets 2
Loading

v0.4

13 Mar 04:28
@jhamman jhamman
v0.4
b82532f
Compare
Choose a tag to compare
v0.4

v0.4.0 (12 March 2017)

This release contains a number of new features and bug fixes.

New Features

  • GARD documentation is now hosted on Read-the-docs (#3, #13, #19)
  • Continuous integration testing on Travis-CI (#4)
  • Add timezone option to handle different timezones in the training/prediction data for the GEFS option (#18)
  • Default transform was changed to "No Transform" (#21)
  • New option for using training data mean/std to normalize prediction data. This is particularly useful when the prediction data record is quite short. (#22)
  • Improved minimum value handling in normalization (#25)
  • New option for passing the prediction straight through GARD without any statistical modification (#27)
  • Added new interactive option to allow users to specify what command line behavior they want (#31)
  • Improvements to output files, variable names are now sensible (#33)
  • New agg_method option (GEFS Only) to allow the method (sum or mean) that GARD will aggregate over the time indices in the prediction data (#35)
  • Weighted averaging (GEFS Only) #37
  • Allow for users to cache/reuse coefficients from the pure regression option (#39)
  • Experimental post-processing utilities inside of GARD (#49)
  • Experimental QQ-Normal transform (#51)

Bug Fixes

  • Fix to time precision / calendar gain (#14)
  • Fix to pure regression option when DEBUG=FALSE (#24)
Assets 2
Loading