ISIMIP3b: GFDL-ESM4 bias-adjusted tasmax and tasmin data replaced

Posted by Stefan Lange on May 29, 2020

A bug in ISIMIP3BASD v2.3 caused bad bias adjustment results for the diurnal temperature range (and hence tasmax and tasmin) in some grid cells and calendar months. Whenever this happened, the entire application period was affected. The bug was fixed in ISIMIP3BASD v2.3.1, which was used to re-do the bias adjustment of tasmax and tasmin. The old data were replaced by the new data on May 29, 2020.

The plot below shows differences between the old and the new data. In particular, it shows the annual global maximum absolute difference between data bias-adjusted with v2.3 and v2.3.1 of ISIMIP3BASD. Nonzero values in these time series indicate which files were affected by the bug. For these files, the particular grid cells that where affected are shown here for picontrol (1701-1750) and here for ssp585 (2071-2100).