transient modflow in gms

[ellibellike]

I would like to calibrate a transient model. I have imported transient well monitoring data. Is it possible to compute output of wells at observation points like with the steady state observation points? Then it would be possible to see residual data. Is it possible to assign observation data to a stess period ? The different stress periods in this model are steady state.

thanks for your help

ellen

[Sean Czarniecki]

The quick answer is "yes." The long answer requires you to go through some of the tutorials which I'm pretty sure used to cover this (I haven't done it for a few years). I don't think you necessarily tell it which stress period it is in, but if the results line up with the output time steps/stress periods, that's what it will use. Therefore, you might have to move some of your observation data to new dummy dates. I'm really scraping the bottom of my brain on this one, so it could be much simpler than I am thinking.

[Alan Lemon]

I would like to calibrate a transient model. I have imported transient well monitoring data. Is it possible to compute output of wells at observation points like with the steady state observation points? Then it would be possible to see residual data. Is it possible to assign observation data to a stess period ? The different stress periods in this model are steady state.

thanks for your help

ellen

Look at the "Managing Transient Data" tutorial: ftp://pubftp.ems-i.com/download/GMS/GMS%2...ient%20Data.pdf

This gives an example of a model with transient observation well data.

You can see the computed vs observed over time for a particular well by using the time series plot. This is one of the things that the tutorial shows.

[Key]

Hi, I too am trying to calibrate a transient model, have read the suggested tutorial but still have a question.

When using computed versus observed data plot it only seems to plot the very first observed data point for each location, irrespective of the time (some are more then 30 stress periods apart).

Is there a way to plot the computed versus observed for a particular stress period? If not is there a way to plot more then just the first observed data point for a location?

Thanks

[Ron Kent]

Hi, I too am trying to calibrate a transient model, have read the suggested tutorial but still have a question.

When using computed versus observed data plot it only seems to plot the very first observed data point for each location, irrespective of the time (some are more then 30 stress periods apart).

Is there a way to plot the computed versus observed for a particular stress period? If not is there a way to plot more then just the first observed data point for a location?

Thanks

The same problem doesn't happen for me. When I have the head data set selected and move through the time steps, the computed and observed values both change in the Computed vs. Observed Values plot. The observed data points shown agree with the values entered through the Map module for the different time steps. I think I remember a bug similar to what you're describing a while back (probably about a year ago). You might want to download the latest update of GMS by going to www.ems-i.com, clicking the download link, logging in (you might have to create an account first), and then selecting the GMS executable update to download. This might solve your problem.

Another plot that might be helpful is the Time Series plot. In the Plot Wizard, select Time Series as the plot type, hit next, check the "Observed Values" box, and select the points you want to show. It will draw computed values vs. time and observed values vs. time for the locations selected. The problem is that this plot can get pretty crowded if you select too many locations.