Dear authors,

Please find attached an annotated pdf with comments.

A main point of improvement I see is that the manuscript can (should?) account for sampling uncertainty in the KGE scores. For comparative analysis such as this (where differences in KGE scores between various options are used to decide which of the options is better), it is important to get some idea of the uncertainty associated with the scores themselves. An R tool exists that makes estimating these uncertainties for NSE and KGE straightforward from existing time series of observations and simulations (Clark et al., 2021) .

More generally, I think the mansucript can be clarified here and there (see comments in the pdf) and possibly streamlined a bit (though I'm unsure what to scrap - it just seems to have a large number of figures).

Kind regards,

Wouter Knoben

Dear authors,

Please see the enclosed pdf for minor comments.

Kind regards

Trine J Hegdahl