Jump to content
GMS, SMS, and WMS User Forum


  • Content Count

  • Joined

  • Last visited

Community Reputation

0 Neutral

About Michal

  • Rank
    Senior Member

Profile Information

  • Location
    Czech Republic

Recent Profile Visitors

1,586 profile views
  1. Hi Lalith, this is because you used layer range. If you use Screen Top/Bot GMS will use different algorithm and the amount specified will be total for entire well. Or you could just leave it as Layer range and set the Q value as 1/3 of the total rate (if the well is set in 3 layers). I belive this is intentional GMS feature.
  2. Hi, well, I think this feature that is not working properly. Here is an excerpt from MT3D manual, that describes how it handles units: "The MT3DMS code uses any consistent units for input and output variables. In the input file to the Basic Transport Package, the user decides the units for time, length, and mass. Then, any input variable or constant should be entered in units consistent with the three basic units. ..." [chapter 6.2 of Zheng, Wang (1998): MT3DMS A modular three-dimensional multispecies transport model for simulation of advection, dispersion and chemical reaction
  3. Hi, I agree it is poorly documented. I had to dig into the source codes to figure it out. It seems to be an excess infiltration volume rejected by the model. It can be routed to other packages as described in the ModelMuse documentation: "If the applied infiltration rate is larger than the vertical hydraulic conductivity, the actual infiltration will be reduced to the vertical hydraulic conductivity. Any remaining potential infiltration will be available to the MVR package as rejected infiltration." So I guess if you lower your infiltration rate this term should approach zero.
  4. Hi Bruce, it is hard to tell. I suggest you to compare the HK and Sy arrays of the HDF5 and native MF version of your model. See if they are different. They should be exactly same. You could read the HDF file directly using Python for instance, see the discussion here. Good luck Michal
  5. Here is the relationship from the manual:
  6. Hi, I think what you are trying to achieve could be done only aproximately, because you only have to assume what the head in the model will be. Anyway, for a given cell Qb = cb (hb - h), where Qb is the flow into the cell from the source cb is the conductance between external source and the cell hb is the head assigned to the external source and h is the model calculated head in the reference cell You problably have measured hb, so I would calculate cb this way: cb = Qb / (hb-h) But as I said, you can only assume what the value of h will be. Maybe calibrating the m
  7. Hi, it is actually quite simple. Each stress period is divided into one or more time steps. The time steps within a stress period does not have to be the same length, but can increase exponentialy in length in order to capture system response to change in stress more accurately. So by specifying a multiplier f, each following step is f multiple of the previous step. So if you have stress period of lets say 10 days and you want to solve it in 3 time steps with 4.0 as multiplier, modflow would calculate the following step lengths: {0.47619, 1.90476, 7.61905} It is not so hard to figure t
  8. There is a trick that could be used with MODFLOW. One could think of the Z coordinate as Y and model it as single layer model. But this is off the table for SEAWAT due to gravity vector orientation.
  9. Hi, I think you could use the conceptual model approach. You need to think of your cross section model as a normal layered aquifer model, that has only a single row. Therefore everything has to be treated layer-wise. You need to follow the MODFLOW logic, there is no special case for cross section models. And you obviously need to add Y coordinate to your elevation dataset.
  10. Hi, well the message desribes the issue. There is only one point in the scatter point dataset you interpolate from. I would double check that the scatter points do not have this error.
  11. Hi Jonathan, I was dealing with similar issue some time ago. There is not a satisfying way how to do it in GMS. The solution depends on the complexity of the task, but it involves grid level approach. You can isolate cells by material, select boundary cells by intersecting with boundary Arc (from Top view, Orto off) and adding a Sink/Source term. In some situtions it may be convenient to map BC form conceptual model to all layers and delete from grid where appropriate. This works if the boundaries do not overlap. Or you can modify the Package outside GMS with some script.
  12. Hi Rachel, The problem is probably that the concentration observation file is not generated after your model finishes. Therefore PEST complains that it cannot find it. The reasons for this could be many. I suggest you to study how model equivalent concentration observations are generated with VM. Alternatively, you could do it manualy with MOD2OBS program from PEST utility suit for instance. When not sure what to do, read PEST manual.
  13. Hi, I think that what you describe could be achieved by general head boundary. Without specifying head at the boundary modflow cannot calculate groundwater gradient between the cell and the external region. So I think the problem is unsolvable the way you try to aproach it. I suggest using GHB package. If you have no information about heads outside the model domain, you could for instace specify GHB head the same as some initial model result head and together with lower conductance the head in the boundary condition cells will vary based on pumping/injection in the model domain. For higher con
  14. Hello, I think that MTD3MS was released long before MODFLOW-NWT. MODFLOW-NWT does not change IBOUND value to 0 if cell goes dry and this is probably the cause of problems you are facing. I think MT3D-USGS has improved handling of such situation: MT3D-USGS also includes the capability to route a solute through dry cells that may occur in the Newton-Raphson formulation of MODFLOW (that is, MODFLOW-NWT).
  15. Hi, I think it is not directly posible through MODFLOW, but maybe you could create the second drawdown dataset in postprocessing using GMS data calculator. Just compute difference in head between the first timestep of the first simulation and the transient head dataset of the second sim. That should be the drawdown you are looking for.
  • Create New...