Nash number is varying even for regions whose parameters are fixed?

[Afonso Augusto Magalhães de Araujo]

Hello everybody!

I am calibrating the basin illustrated in fig_1-stations (attached), step by step, from upstream to downstream, freezing the parameters for the HRUs/AQUs.

For each new FLU channel/station to be calibrated, I create a new project, insert the upstream calibrated/frozen parameter sets, and a variable default parameter set for the new channel.

Before proceeding to automatic calibration, I run manual calibration to ensure that the parameters have been correctly entered and that the NSE (and other) indicators are being reproduced correctly. This is what happens in most cases (as expected?), but curiously/inexplicably, not always.

After calibrating 7 stations with relative success, the following situation is occurring at the eighth and final station:

I created a new project, inserting one by one the 7 sets of calibrated and frozen parameters. I ran the automatic calibration and the NSE indicators (and others) were reproduced with very small differences, which is already curious (fig_2-Before). After inserting the new initial variable parameter set to be calibrated for the new channel (station) and running the manual calibration again, there was an unexpected variation affecting the upstream results (fig_3-after).

From this, some questions arise:

1. Is it possible that, even with fixed parameters, there could be variations in the upstream calibration results? In my experience with SWAT_CUP, this did not occur;

2. Initially, when entering the calibrated parameters for only the first seven stations and running the automatic calibration, I obtained an NSE of 0.92 for the eighth channel/station even before automatic calibration. If I were satisfied with this result, how could I access the values ​​of these parameters? How could I determine which HRUs/AQUs they are affecting and freeze them? Is this possible?

3.What could be the reason why, when entering the new set of parameters to be calibrated for the last channel/station, the results for the upstream channels/stations vary?

I would greatly appreciate any help.