Post-processing plots vs. *.out column data

The HP1 code incorporates modules simulating
(1) transient water flow in variably-saturated media,
(2) transport of multiple components, and
(3) mixed equilibrium/kinetic geochemical reactions.
Post Reply
S_Jessen
Posts: 8
Joined: Thu Mar 04, 2010 6:59 am
Location: Denmark
Contact:

Post-processing plots vs. *.out column data

Post by S_Jessen » Wed May 15, 2013 9:09 am

Dear Diederik

I just ran through example 1 in the well written HP1 Tutorial II. It worked fine, of course.

However, example 1 is fully saturated, so I then changed a bit to check out the unsaturated capabilities etc. Here's what I did:
1. Loaded the freshly downloaded HP1-1 example file ran it in my freshly installed Hydrus 4.16.0080. Post-processing plots for "Observation Points" looks fine/consistent/reliable. The plots for "Water Flow - Boundary Fluxes and Heads" shows "not enough data" in the plot window. I recon this is due to the unchecked "Water Flow" in "Main Processes" menu.

2a. To try the same but for unsaturaed steady state conditions, I then opened "Soil Profile - Summary" and clicked "Set Initial Conditions Equal to Field Capacity" (results in h = -107.8 cm thoughout the profile). Then OK, save and run: Post-processing plots now have errors. In "Observation Points", for instance, "Water content" and "Total_H" are negative and time dependent. The Time vs. "Concentration: Cl"-plot shows signs of numerical lack of convergence, jumping apbruptly up down; more about Cl in 2b below:

2b. The ts_tot_Cl.plt opened in GNUplot is empty. However a god sign with respect to model performance is that the nod_inf_chem.out shows nicely decreasing Cl values with time, at all times. Also when having the "mass_H2O" included in the nod_inf_chem.out-file (i.e., checking "Water" in "HP1 Print Information" menu), this parameters shows to be constant (=0.20; i.e., field capasity, I assume) with time and depth.

3. For fun, I then unchecked "Solute Transport" in Main Processes menu, hence leaving none of the boxes checked. Save and run: Results in Post-Processing plots are fine and consistent, so Hydrus without PHREEQC seems to work nicely for steady state unsaturated flow.

For now, I'll trust the chemical results available in the nod_inf_chem.out-file (except if you tell me not to), but here, I'd just like to point out these experienced inconsistencies in the othervise very clever coupling of Hydrus with Phreeqc, which I truely love :-)

Best regards
Soren

***********
Soren Jessen
University of Copenhagen
Water Resources Research Group
Dept. of Geoscience and Natural Management

diederik
Posts: 51
Joined: Fri Mar 10, 2006 7:57 am
Location: Belgium

Re: Post-processing plots vs. *.out column data

Post by diederik » Wed May 15, 2013 7:54 pm

Hello,

1. Indeed, when the 'water flow' option is unchecked, no information with respect to water flow is outputted in HP1. We might change that in a next release

2. This is an output format-writting bug which was already identified earlier and is correct in the next version: Very small concentrations (< 1E-99) are wrongly written as e.g. .127-207 instead of .127E-207. When the HYDRUS-GUI makes the graph from the output files (based on obs_node.out), this is wrongly interpreted and plots are completely messed up. When you look at the obs_nod_chemxx.out files, concentrations are plotted correctly there. If you open the pd_tot_Cl.plt file, you see nice Cl profiles but limited to the few first centimeters. I do not think the ts_tot_cl.plt file is empty, but Cl concentrations are all equal to 0 (and the scale is not optimal). It will look 'not empty' if you have an observation node at the first.


So, I am quit confident that you can trust the results of the *chem.out output files :D and we will solve this inconsistency in the next release to avoid confusion.

Diederik

S_Jessen
Posts: 8
Joined: Thu Mar 04, 2010 6:59 am
Location: Denmark
Contact:

Re: Post-processing plots vs. *.out column data

Post by S_Jessen » Thu May 16, 2013 8:17 am

Sounds great to me, thanks!

Best regards
Soren

Post Reply