| Author |
 Topic  |
|
|
jand
Junior Member
4 Posts |
 Posted - 03/04/2010 : 19:23:29
|
Dear Jirka/Mirek,
Can Hydrus simulate seepage through a boundary with an atmospheric BC?
I assumed that when the groundwater level in a variably saturated domain (XZ) rises and touches the soil surface in depressions in the landscape, the algorithm for the atmospheric BC switches the node to a constant head (h=0) and seepage can start. However the solution seems to oscillate (h_Mean.out indicates that hAtm jumps from -0.586 to 16m in a very small timestep <0.001day), and the program stops. This happens exactly at the time when the simulated phreatic level starts touching the soil surface at a low point in the flow domain.
Our XZ flow domain represents a cross-section through phreatic aquifer with vadose zone on top. The cross-section represents a transect perpendicular to a stream. The top boundary is a very gently sloping soil surface fully covered with grass (no evaporation, only transpiration), which I assign an atmospheric BC (receiving rain). The right boundary is the water divide, hence assigned a no-flow BC. The bottom is the base of the aquifer and equally gets a no-flow BC. On the left, the domain runs up to a stream. The nodes representing river bed and river bank receive a variable head condition (with the value of the lowest node given in ATHMOSPH.IN; water level assumed constant in time). The left boundary below the middle of the river bed is also assigned a no-flow BC (symmetry). The area is a wetland, and through the definition of the domain and the Ksat value, the model indeed simulates that the water table rises during prolonged rainfall periods until it touches the soil surface in the lower parts of the landscape.
When we define the entire upper boundary as a seepage face, AND also specify as BC option that we want to apply athmospheric boundary conditions to non-active seepage faces, the model runs fine and seepage does occur. Is that the correct upper BC to chose for our problem? That option however has the disadvantage that the file cum_Q.out does not give any output on infiltration or runoff. Infiltration, seepage, runoff seems to be all lumped into the CumQS variable (seepage). Correct? Any solution to that?
Kind regards,
Jan
|
|
|
Jirka
Moderator
USA
1788 Posts |
Posted - 03/04/2010 : 19:42:40
|
If you select only Atmospheric BC, then the code only reduces fluxes when limiting pressure heads are reached (for infiltration or for evaporation). It does not start doing any other things, such as creating seepage faces.
Yes, you need to use an option which allows you to apply atmospheric fluxes on the nonactive seepage face. It is currect that in this case, fluxes are reported as fluxes across the seepage face. You could possibly look into the Boundary.out file, which gives info on all boundary nodes. Look for all nodes with the kode=2 (active seepage face) and Kode=-2 (inactive seepage face). By integrating (adding) Q values for these nodes, you can get flux through each part of the seepage face boundary.
Jirka |
 |
|
|
jand
Junior Member
4 Posts |
Posted - 03/04/2010 : 19:48:16
|
Thanks!
Jan |
|
|
| |
Topic |
|
|
|
Atmospheric BC and seepage
