Extracting depth to water table

[MABOSSANI AKARA]

Hello All,

I am using Parflow to simulate temporal changes in water table.

I am using an orthogonal grid with solids to define the active model domain.

I was able to run the model, but I am having issues extracting the depth to water table. The saturation file shows that the position of the water table is correct. 

But when I try extracting the depth to water using  "hydro.calculate_water_table_depth(pressure, saturation, dz)", the output does not look right.

I have attached the plotting ('plot_wt_og.py.txt') scripts and will appreciate any suggestions.

[Georgios Artavanis]

Hello,

Thanks for your message.

Can you please provide some more information about your run? I see in the pfidb file that the dimensions of your domain are NXxNYxNZ = 24x4x610. Are you trying to plot wtd across one of these dimensions? I think the problem might lie in the line

ax.plot(list(wtd[0, ...])[2],'b', lw=0.5, label='water table')

Can you share the shape of your wtd array and what you are trying to plot in this line? My guess is that the list is unnecessary and that it might cause some bug in the plot. Instead, I would use NumPy indexing for clarity.

In addition, the same can be said about this line:

ax.plot(list(z0-wtd[0, ...])[2],'b', lw=0.5)

I'm not sure what the correct formulation is here (whether you should subtract from your elevation or not), but the list might again introduce a bug. if you can share the intended behavior in the plot, we can do more debugging.

Best,

George

[MABOSSANI AKARA]

Hello George,

Thanks for your response. As you correctly pointed out, my domain grid is 24x4x610.

The wtd array is  4x24 (NYxNX) and I am trying to plot the water table along the x-direction.

As suggested I used  "ax.plot(wtd[0,:],'b', lw=0.5, label='water table')" but the issue remains.

Below is the expected plot:

I am not sure if the inactive cells above the land surface and below the bottom surface are contributing to my issue.

Thanks

[Georgios Artavanis]

Hello,

Can you please share how you define your domain and how you constructed your solid file (pfsol)? It could be, as you said, that inactive cells are contributing to the issues. Also, could you simplify the plotting by plotting just the wtd from that step over the entire domain (as a 2D plot, for example with plt.imshow())?

We also noticed that DX, DY are significantly larger than DZ, and we would like to confirm that this is intended to be so for this domain.

Thanks,

George

[MABOSSANI AKARA]

I used the 'pfpatchysolid' tool to prepare the solid files (see create_slds_otg.tcl).

I prepared one solid for the entire domain and 5 other solids files to assign hydraulic parameters.

The python script used to run Parflow is attached ("s8_pf_og2.py").

Below is a plot of the water table.

I noticed that for most of the domain (in previous figure), the "parflow.tools.hydrology.calculate_water_table_depth" function returns NaN (4 by 24 array).

I thought a fine discretization is required to better map Parflow solids onto the computational grid; this is the reason for the small dZ.

Please let me know if my understanding of how solids are mapped onto the computational grid is accurate.

Thanks.