Time step smaller than 3600 sec

[Mayra Ishikawa]

Hello everyone,

I would like to know if anyone had used a time step smaller than 3600 sec (which is the default).

From other studies I could not find anyone who changed the sub daily dt to less than 1 hour. I wonder why people usually don't do that since it is possible to make this change over the configuration options.

There is one simulation that I'm working with and we did all the set up and calibration with the default (3600 sec) and it looked good (see figure 1 attached). We eventually changed the dt to 1 sec, because we were comparing it to other models that had smaller dt. 

Initially we thought that it would not have great changes, since only increasing the dt could make models numerically unstable. But for our surprise there were significant changes, specially more mixing (figure 2). 

According to the GLM publication at gmd, numerical diffusion should not be a problem.

Any hints on why results are so different?

Our boundary conditions are in a mix of daily temporal resolution (inflows and outflows discharge and temperature) and hourly temporal resolution (meteorological conditions). But I believe this should not make a big difference, since the model would then interpolate the data. Is it right?

Kind regards,

Mayra

Figure 1 - dt = 3600 sec

Figure 2 - dt = 1 sec