Jump to content
GMS, SMS, and WMS User Forum

Rob Virtue

  • Posts

  • Joined

  • Last visited

Everything posted by Rob Virtue

  1. Given the increasing use of Null-Space Monte Carlo analyses in models, it is a significant gap that they cannot be done for MODFLOW-USG models. Any chance this could be added?
  2. If you have multiple coverages for flow budget zones, then map a visible zone coverage to MODFLOW, when you run the USGS ZONEBUDGET tool it will pick up all zones, whether active or not, even if they overlap. This is much like the bug with grid activation coverages where all "activation" areas are mapped, including inactive coverages. The worst case is where you have zones with the same ID number in more than one coverage and there is no way to tell which is the "real" zone. Assuming you don't want to delete the alternative zone coverages, it can be fixed by de-selecting "Zone budget ID" in the "Aerial Properties" box for all inactive zone coverages. This bug cost me a lot of sleep and hair over the weekend!
  3. The ability to export a model to a 3D PDF (as done by various GIS packages, SURFER etc) would be fantastic for non-modellers to visualise the output rather than relying on 2D views or animations.
  4. If you have a river stage below bed level, you will get a warning but the model may still run. Unfortunately, the result will be garbage, sometimes spectacularly so. This is because the package includes the calculation stage elevation - bed elevation to calculate gradient, so then the value becomes negative. It could be solved by inserting a line of code that sets flow to zero if stage is at or below bed level, but for some reason that has never been implemented by USGS (perhaps I'm missing some consequence of doing so). I once got someone to modify an earlier version of the MODFLOW source code to include this and it seemed to worked well. Assuming you can't do that, then all you can do is make sure your river bed TIN matches your model cell grid (turn 2D grid into TIN) and either truncate your stage elevations so they are no lower than the bed elevation or set the bed elevation to below the lowest stage elevation. Then make sure the river cells are assigned to layer corresponding to the lowest level. Note that for transient TINs the layer assignment is based on the first time step, not the lowest (or at least it used to be) so make sure you have the lowest values in the first time step, even if you have to "turn of' the river with a very low conductance in the first (short) time stem. Sorry there is no easy solution.
  5. Is it possible in NWT for the hydraulic characteristic of an HFB to change over time? If so it would be useful to have this as a transient option, much like GHB conductance, to simulate the installation of HFBs over time and their degradation long-term.
  6. Arto, hydraulic conductivity determines how much volume flows through fro a given aquifer for a given area an gradient. Imagine a block of aquifer and a block of water being extruded out of the side as it flows. If the aquifer was all space (porosity of 1) a particle of water would flow through the aquifer at the same rate is is extruded out the side. If the porosity is only half (0.5) and everything else remains the same, the velocity of a particle in water in the aquifer has to flow twice as fast to keep the block of water moving out. If your porosity is 0.01 for the same K, it has to move 100 times as fast. This is expressed in the Darcy flow velocity equation V=K*(dh/dl)/Pe where V = velocity K= hydraulic conductivity dh/dl = hydraulic gradient Pe = effective porosity. Particle flow velocity is inversely proportional to effective porosity, if all other parameters remain the same. This may seem counter intuitive, because you are thinking that less pore space means lower flow, but remember, you are keeping hydraulic conductivity the same, so the volumetric water flow stays the same.
  7. What conductance value are you using for the drain and what are the Kh & Kv of the high conductivity layers? A higher drain conductance is required to fully drain a higher conductivity layer.
  8. How are the drains mapped? If they are from arcs, the elevations will not necessarily follow the variation in topography, especially if topography is irregular and node spacing wide relative to the grid. If as a polygon, for the same reason, elevations may not match if topography is irregular and the grid and raster resolution are not similar. Solutions are : Use the automatic layer range allocation with "Auto-assign BC to one cell", so they can go into the layer below if below the top layer (if that is conceptually acceptable) If using arcs, redistribute vertices more finely along the arc, to a similar spacing to the grid, convert them to nodes so it can have more variation over distance. Simplest and most robust option is to use a Seepage Face rather than Drain, so that the drain elevation in the grid is set relative to the upper cell top elevation. In your case set the “elevation offset” set to -1m.
  9. Can you please add the new Australian mapping datum, GDA2020, to the projection options.
  10. In Windows 7 (and presumably other versions) , it depends on the font size selected under Control Panel\All Control Panel Items\Display.
  11. Dunaj, Check under HELP>REGISTER> that you have the TIN module enabled.
  12. I used to get this a lot but it's been a while. If it is a transient model, make sure there are no observations in the first time step (or maybe first stress period?) . This can stop the model running or the head file saving properly.
  13. Rob Virtue

    ET surface

    Kirk, We had the same issue recently (18 September 2015), with the option for "2D Dataset -> Array" being greyed out, and were given this work-around from Jack at Aquaveo support. It seems to have worked. "Currently the option to use the "2D Dataset -> Array" command for U-Grids is not available, however I believe that the developers are working to implement it in future versions of GMS. I've got a possible work-around for you. This assumes that your topographic surface can be represented as a 2D scatter set or TIN. In this case right-click on your U-Grid and select 3D->2D, and then select layer 1 so as to make a 2D U-Grid of the top layer of your 3D U-Grid. This ensures that the cell IDs of layer 1 in your 3D U-Grid line up with the cell IDs of the 2D U-Grid. Now select the 2D U-Grid to ensure that it is the active U-Grid, and select the topographic surface data that you have and convert it to a 2D scatter set. Then interpolate that 2D scatter set to the 2D U-Grid. Now you will right-click on the interpolated dataset on the 2D U-Grid, and select "View Values". Copy the column of f values here, click ok to exit that dialog and bring up the dialog for the EVT package on the 3D U-Grid. Now paste in those values to the array for the EVT package. You can check to make sure that the cell IDs are the same and line up all right. If you have trouble doing this, please let us know."
  14. Mina, I suspect you have specified head cells with the head elevation set below the base of the cell, which causes a terminal error. Use the automatic layer assignment option to ensure the elevations are set at the uppermost cell ("Auto assign to one cell") or the uppermost and all underlying cells ("Auto-assign including lower cells") where the elevation is above the base of the cell. If you are using transient heads, be aware that automatic layer assignment seems to use the elevations from the first stress period.
  15. Mgozdor, This occurs if you have the "use to define model boundary (active area)" box checked in ANY coverage in addition to your activation coverage. This is a common problem at the moment, because if you select "layer range" in a coverage, it automatically checks the "..active area)" box. A bit annoying really. It will cause the activation errors whether you map the coverage to the model or not, so make sure you check ALL of your coverage.
  16. Rob Virtue


    I am having the same problem (GMS 10.0.6). It appears that PEST is looking for flow observations for DRN and CHD boundaries, even though there are no coverages with flow observations in the current model.
  17. If you have a high conductance value for the general head boundary relative to the cell K, it will act as a variable head cell as there is not a significant restriction to flow. If you drop the conductance to a very low value relative to K (conductance can be varied over time just like head) flow is restricted so that the is negligable exchange to and from the GHB term. By varying the conductance, the GHB can be used like adding or removing a constant/variable head boundary, without the need to stop and start the model.
  18. Another option would be to use a general head boundary instead, with heads as you are using for the constant head boundary, and use the conductance values to turn it "on" and "off".
  19. I was editing a large model and added one new scatter point, which caused GMS to crash. I could then not reopen the model. After the panic attack subsided, I noticed that my hard drive was full. After deleting or moving about 3 GB of files off my hard drive the model opened and edited OK for a while and then crashed again. I noticed the hard drive was full again and did some hunting around and found that there was about 30 GB of world imagery files in \temp. Two Lesson: - make sure the "online maps" box not checked in GIS once the image is downloaded, to prevent is downloading again every time the view changes; and - regularly clear the cache (under "advanced" options in "get on line maps") or delete temp files. Panic over and back to the model. And backing up the model.
  20. Are transient TINS and PHT3D in the recent release of GMS9.1.1?
  21. It would be handy to be able to have a small comment field attached to various data bases, coverages, scatter point stes etc to explain their use. For example there may be multiple recharge coverages for different climates or development scenarios. While this can be adressed to some extent with the coverage/set name, when more detail is required this would result in a very long and complicated names.
  22. Alan, I created the first layer coverage with materials and various boundary conditions (GHB, RCH, DRN) assigned to layer 1-1. I then copied the coverage to produce the materials + colour polygons for the lower layers (2-4) and unclicked all the boundary/source/aerial property conditions, so that the copies only mapped colour and materials. Despite changing the materials and colours in the new coverages, when they were all mapped only the layer 1 materials applied. If I mapped the layers individually all was well, until I mapped the top layer. The problem was fixed by creating another copy of the primary, layer 1 coverage, for layer 1 materials and colour only. Thus I ended up with 5 coverages: boundary conditions for layer 1 (all other layers had no-flow external boundaries and no internal boundaries); then materials and colour for layers 1-4. It is not a problem I have ever noticed before, although I do not often use coverages to apply materials to layers.
  23. Using 8.3.5 (and presumably older versions). If you are using material coverages for different layers, do not combine other attributes in the material coverage. If you include recharge in one of the material layers (typically layer 1) it will copy the material along with recharge from that coverage over all layers. If only materials (and colour) are used in the coverages, they map properly to the appropriate layers.
  • Create New...