Jump to content
GMS, SMS, and WMS User Forum

Sean Czarniecki

GMS Experts
  • Content count

  • Joined

  • Last visited

  • Days Won


Everything posted by Sean Czarniecki

  1. Just for kicks, you might try lowering your specified head to see if you can achieve a reasonable calibration. Have you accounted for evapotranspiration? Could your recharge be too high? Are your calibration measurements being influenced by any tidal effects (you said "coast," but are running steady state, which is an average condition....perhaps your measurements were taken under a lower tidal influence)? Just some things to consider.
  2. Water level change with recharge

    Not knowing what your boundary conditions are, I'll assume that the cell goes dry in an area which isn't impacted by boundary conditions (even if no-flow). Your starting head may be too different than what the final elevation is going to be. The calculations will do an interation which may come to a solution below the bottom of the cell (think of the iterative process as a wave where the model goes up and down at smaller amplitudes until it gets to the solution). If the low value is below the bottom of the cell and there is no rewetting, the cell is dry and won't rewet. You might have to try starting heads that are actually lower to be closer to the final solution. Just my initial thinking.
  3. Modflow 2000-transient model result

    You've provided some good information. However, I can only provide you my thoughts based on what I see. It is unclear to me which layer(s) you are pumping in. Based on the K of layer 1, that wouldn't be a great layer to withdraw water from, compared to Layer 2. It is not shocking to see Layer 1 cells drying out around the wells, based on the low K value. I'm not seeing anything really wrong with your head results. They will not be exactly the same (which is how you have upward/downward gradients in an aquifer).
  4. Modflow 2000-transient model result

    When you have a confined layer, it typically doesn't dry out. Therefore, if it *should* dry out, there is a good chance that you are going to see unrealistic heads in the model for that layer. You might want to make Layer 1 be unconfined and see how that goes.....or the Layer 2 head may be more appropriate to use (depending on what you are trying to show and how close it is to Layer 1 in the other cells).
  5. Tracking particles in Modflow-USG doesn't works

    Older versions can still be used if you find a newer one has a bug. I have done that many times in the past....and if you have any issue with the license, getting a temporary one usually isn't a problem.
  6. Cell dimension change more than 50% in .............

    Note that it is just a warning and not an error. The model will still run. Typically, it is preferred to have cell sizes transition from smaller to larger without too big of a jump. Better for numerical calculations. This warning is probably due to your layer thicknesses being more than 50% different....and is not unusual, especially when dealing with lower layers in a model.
  7. Observations in MODFLOW-USG

    Boy, you want us to actually read the documentation? Thanks - I was actually hoping that you could point me to something like that!
  8. Observations in MODFLOW-USG

    At one point during my current project (a transient MODFLOW-USG simulation), I was attempting to create observation points to make some plots. It didn't seem to be working. I know that I've had some minor issues in the past getting the observations to work correctly, so I didn't think too much of it and got the info I needed in another way. However, I found myself with a little time today and figured that I would look into it a bit more. I pulled up an old model where I had a transient simulation to see if I had done anything different (I did not). I then ran my model using MODFLOW-2005 (after removing USG-specific stuff) and even though I didn't converge, the observations points appear to have worked. Are observations not set up to work in MODFLOW-USG?
  9. Potential Bug in CLN Well pkg?

    I had a transient model set up where each of my well flows were going to be the same for each stress period, so I checked the "Use previous" box in the CLN Well Package dialog. When I got my model results, the first stress period had the correct total flow in the output file, but all the other stress periods only showed the flow from the first well. I tested things further by unchecking the "Use previous" box for other stress periods. The correct total flow was shown for those stress periods. I am now running with all stress periods not having the "Use previous" box checked. Just an FYI. This is GMS 10.2.5 build date of May 17, 2017.
  10. Fonts and Icons in 10.3

    I'm sure there is an easy way to change the project explorer fonts and icon sizes in 10.3 (they are much larger than I would like on my screen...much different than 10.2), but I'm not finding it. If anyone knows and can pass on that info, I would appreciate it. Thanks!
  11. Potential Bug in CLN Well pkg?

    Just wanted to let you know that this is still not fixed in the latest build (GMS 10.3.3). I haven't checked the latest build of 10.2.
  12. CLN well head results in binary file

    I would like to see head results at each individual CLN well in my model. They are supposedly output into a binary CLN_HD file....but I can't seem to look at them in GMS. Can anyone help me out here? The CLN_CB file gets pulled in to the model after the run, but the CLN_DD, CLN_HD, and CLN_IB files do not appear to be. This is time-critical - if anyone has a way for me to see these results, I would appreciate it. (note that I'm using GMS 10.3.3 - I don't think that the CLN_CB file even got pulled into GMS 10.2, but I haven't fully looked into it) Thanks!
  13. CLN well head results in binary file

    Right. I have around 200 wells in my CLN network. I need to see the heads at each individual well. These values are output into a binary file. I'm kind of shocked to see that GMS produces a binary file without reading it in afterwards. Do you have a suggested way that I could look at the results by opening the binary file directly? I'm looking at some free software, but it didn't seem to translate well. This raises another question - when the solution contours are produced for display in GMS, do they take the individual CLN wells into account? I have a feeling that they don't, even though that is one of the reasons to use the CLN package along with MODFLOW-USG (to be able to have multiple wells in one cell and not have them act like one well for the entire cell).
  14. problem horizon to solids

    It could be that some shallower boreholes are throwing things off for you if they don't go down deep enough to define your layering. I know that in the past I have made polygons in my cross sections which extend downward from a shallow borehole to help fix this issue. I can't say it will work for your situation - you may just have to either put in a dummy borehole or remove the shallow one.
  15. Model Checker with SFR2 Package

    I second the suggestion, although the MODFLOW output file tells you this immediately. The model crashes and the elevation errors are listed right at the end of the output file.
  16. MODFLOW-USG & CLN cells staying wet

  17. MODFLOW-USG & CLN cells staying wet

    I've been given a MODFLOW-USG model (including use of the CLN package) whereby the upper layer is Convertible Upstream. The model won't converge if the layer is just set at Convertible. Fine. While I'm fairly new to USG (still reviewing the calculations/equations at this point), I can accept that the model needs that layer to be Convertible Upstream. My real issue is that cells in that layer which would normally be "dry" (head below the bottom of the cell), stay active and report back the head (and use in contouring) which is below the bottom of the cell. If the layer were confined, I could understand it, but I'm guessing that there is something that I don't know about "Convertible Upstream" which is keeping these cells active, even though they should be dry. Anyone have insight into this? I don't really see it as a problem (the model results seem correct and water just passes through the cell to get to the lower layer), but rather something I'm curious about, as I'm not used to seeing "dry" cells stay active before except when modeling the layer as confined. Thanks!
  18. SFR Package - Arc interpolation control

    This may end up under the GMS development request forum, but I'll first see if I'm missing something. I'm using the SFR package for the first time. I've taken one of my stream segment arcs and input the upstream and downstream elevations. When I map to MODFLOW, it would appear that there is limited control over the interpolation. This is in a mountainous region, so the terrain does not necessarily stay at a steady gradient. This causes my stream to often be at a higher elevation than reality. From what I can tell, my only option is to break up the segment into many smaller segments. This is quite painful. I had considered if I could do anything manually in the input files, but the SFR package information goes into a binary file. Rather than breaking up the segment, wouldn't it be easier to change the interpolation to work with the vertices and allow the user to just change the vertex elevation? Maybe it does this already, but I don't see it discussed anywhere....and since all my vertices are at a Z elevation of 0 (and the interpolation isn't affected by that), I'm thinking that they are ignored. Any thoughts on making this process easier than breaking up my segments into many smaller segments? I might seriously need to make segments go from one cell to the next in many locations. Thanks!
  19. SFR Package - Arc interpolation control

    Thanks again - I spoke with my GIS guy yesterday and it seems that getting the segment starting and ending elevations wouldn't be that difficult. I'll just have to figure out how best to get them into GMS. Did you do it via cut and paste into the dialog box for the segments, or through a MODFLOW input file? Right now, GMS looks to a binary file for the array. I think I was recently successful in getting an input file to ignore the binary file, but in another case, I was not successful (I don't recall the details on which package it was), so just curious as to which way you went.
  20. SFR Package - Arc interpolation control

    Thanks, Bruce. I totally agree that the SFR2 package is better than drains or rivers when you are actually trying to get the correct water balance in a watershed. That's why we're using it. The issue is that the region is mountainous, so using one segment for a 30,000 m long stretch of stream/river doesn't put the streambed at the right elevation for much of that length. I like your thought of using GIS, but like you said, it would be pretty high-level work.....I may have to go that route though if I don't figure out a better way to do this in GMS. Do we know if there is an upper bound to the number of SFR2 segments allowed? Without accounting for elevation changes, I will have around 100 segments in the drainage basin. If I wanted to be precise in accuracy, I could easily get to 1000 segments (I wouldn't likely if I did it manually, but if I went the GIS route, I would shoot for high accuracy).
  21. SFR Package - Arc interpolation control

    Okay - I see what you are saying. While there is no actual interpolation which applies a stream bottom elevation to a cell, there is an "internal" SFR2 interpolation which sets the stream bottom in the cell for use during calculations (as evidenced by errors which show up in the output file when a stream bottom elevation is below the bottom of a cell...and by the calculated depth of water in a reach [cell] in the istcb2 file). I might still consider putting this in the GMS development feature request if I have time in the near future. Based on what you are saying, I would want Z elevations from the arc's nodes and/or vertices used to generate segment elevations, as that is the item which is painful to put into the dialog boxes (you can easily put in similar hydraulic properties for all of the arcs, but the elevations are all individual entries). I'll think about what might work best.....there may not be a big push to get that done. This is the first model I've worked on with a large number of streams. I'm sure that most modelers don't go this route too often. Thanks again!
  22. MODFLOW-USG & CLN cells staying wet

    Thanks - I read that, but it wasn't as clear as I would have liked it to be, as other versions of MODFLOW have convertible cells, but once they get dry, they stay dry. The text in the manual doesn't make it clear (to me) that it applies to more than just the CLN cells (and neighboring GWF cells). I guess it does! Thanks again for the response!
  23. Importing MODFLOW-USG files

    I have a model that was created in Vistas which I'm attempting to convert over to GMS. Note that the original model did not include nested grids, so the node number should be the cell number. Anyways, I have the grid set up and was able to get the recharge in. However, the source/sink files (drn, riv, str, etc.) use nodes. I attempted to drop the information into the MODFLOW files, but I get the error that it wants to look to the HDF5 file for the data. How can I get around that? I put in a test drain, saved the file (the data apparently went to the h5 file)....but when I tried to comment out the reference to the h5 file and attempted to add/read in a couple more drains in the MODFLOW file, the model still pulled in the data from the h5 file. I thought that was a bit odd. Anyways, I have all of these node-based points which I should be able to read in, but can't figure out the best way to do it. Any thoughts?
  24. MODFLOW-USG: Recharge Mapping

    Just as a follow up based on your concern, the "thin upper layer" can be as thick as you want it to be to avoid that issue you mention. You can design your layers to meet your needs. Could take a little time, but shouldn't be too bad. Try sketching it out on paper first to visualize how to do it. First the way it currently looks, then with new layer divisions and material properties. Just a little imagination and it will work.....