Bias correction

[seyedmohammad19]

hello,
is there a recipe available for calculating bias correction for cmpi6 datasets, using a reference dataset such as Era5?

[valeriupredoi]

@ESMValGroup/userengagementteam please help here 🍺

[LisaBock]

Hi @seyedmohammad19, as I know there is nothing regarding bias correction available in the tool at the moment. Feel free to add it...

[seyedmohammad19]

Bias correction refers to the process of adjusting estimated values from the CMIP6 (historical) dataset to account for known biases when comparing them to reference datasets such as Era5. The objective is to enhance the accuracy of the estimated CMIP6 values.

There are various methods available for bias correction, the simplest one is the linear scaling method (LS) and the delta method, among others.

we do bias correction for next step and that is:
Assuming that the ratio between the historical and reference data remains constant, and applying this ratio to the future data (like SSP585).

To evaluate the effectiveness of bias correction, it is customary to compare the biases between the historical and reference datasets before and after the correction. Metrics like "bias," "RMSE" (Root Mean Square Error), and "MAE" (Mean Absolute Error) prove useful in assessing these biases. By examining these metrics, we can gauge the degree of improvement achieved through the bias correction process. and we can rank models as well.

about datasets:
Monthly dataset is more appropriate than daily dataset due to the high volume of calculations involved.