Ensemble bias correction using quantile mapping

Bias correction method using quantile mapping. The approach uses a cumulative distribution function-transform method of the entire ensemble to ensure the preservation of the internal variability of members.

Article reference - under review

Overview of workflow

Inputs:

-(A) Model: ensamble members of air temperature (divided into batches) generated using the HadAM4 from the UK Met Office Hadley Centre.

-(B) Observed: ERA5 (with the same grid/resolution)

Outputs:

-(1) Model with bias correction applied

-(2) Histogram of data distribution before and after bias correction in comparison with ERA5

Install dependencies (use Python3)

pip install -r requirements.txt

A detailed description of data and methods are included in the article. A summary of the workflow is described in the next figure.