Warning re Inconsistent units in GMS

[woodward]

As you may know, GMS provides a dialog box where the user may select the dimensional units to be used in their project. According to the GMS wiki,

http://www.xmswiki.com/xms/GMS:Units

GMS:Units

When building a ground water model, it is important to ensure that consistent units are used when entering model parameters. To simplify the management of model units, the user can define the units for length, time, mass, force and concentration in the Units dialog. A units label is placed next to each of the input fields in all the model dialogs in GMS where the units are known. For example, the units for hydraulic conductivity are length / time. If the length units are defined as "m" (meters) and the time units are defined as "d" (days) in the Units dialog, then the units string next to the hydraulic conductivity input field would be "m/d".

GMS does not convert quantities from one system of units to another. GMS only displays the chosen units to help the user make sure they are being consistent.

However this is not quite true, as GMS calculates the value of the gravitational constant g appropriare for the units selected. Also, I understand the new version of GMS also uses the selected length units in its FEMWATER van Genuchten curve generator. I do not know where else these units may be used, but the wiki is certainly not correct that they are for display purposes only.

I noticed that the user can select concentration units that are inconsistent with the mass and length units. In my project I selected m for length, kg for mass, and mg/L for concentration. But given the length and mass scales, concentration should be M/L^3 = kg/m^3 = g/L. I found that this caused me some confusion regarding the units of the distribution coefficient Kd; I was not sure whether it was to be specified in mass-length units [m^3/kg] or concentration units [L/kg]. Going back to the FEMWATER documentation, I think Kd needs to be in the inverse units to bulk density rhob, so in mass-length units. However, if a Langmuir isotherm is used, K should be in inverse conc units. The Freundlich isotherm is dimensionally inconsistent anyway.

I think it is propably safest to always use concentration units that are in the same units as the mass-length scale.

[Michael Kennard]

Thanks for pointing this out. I modified the wiki to address the points you raised. Check it out and feel free to modify.

[woodward]

Thanks for pointing this out. I modified the wiki to address the points you raised. Check it out and feel free to modify.

I was wondering if there are there any other places in GMS where the units actually do matter.

Also there are some dialog boxes where GMS does not provide the units (e.g. FEMWATER materials dialog).

On a related note, I sometimes have trouble with the Fluid Properties dialog box not allowing me to enter the visc and comp values.

Thanks

Edited January 20, 2010 by woodward

[Michael Kennard]

I was wondering if there are there any other places in GMS where the units actually do matter.

I found one more and added it to the list on the wiki.

Also there are some dialog boxes where GMS does not provide the units (e.g. FEMWATER materials dialog).

I've put it on our list to add units to this dialog.

On a related note, I sometimes have trouble with the Fluid Properties dialog box not allowing me to enter the visc and comp values.

I'm looking into this.

[Michael Kennard]

On a related note, I sometimes have trouble with the Fluid Properties dialog box not allowing me to enter the visc and comp values.

It looks like compressibility is only editable if the model is transient, and viscosity is only editable if the model is either transient or you're doing permeability and not hydraulic conductivity. The FEMWATER manual doesn't say anything about this. I can see in the FEMWATER code where viscosity is only used if you're doing permeability, but it's hard to tell how the transientness affects it. I suspect someone figured this out at some point and we added it to the GMS interface thinking we were simplifying the interface and helping users.

[woodward]

It looks like compressibility is only editable if the model is transient, and viscosity is only editable if the model is either transient or you're doing permeability and not hydraulic conductivity. The FEMWATER manual doesn't say anything about this. I can see in the FEMWATER code where viscosity is only used if you're doing permeability, but it's hard to tell how the transientness affects it. I suspect someone figured this out at some point and we added it to the GMS interface thinking we were simplifying the interface and helping users.

If users are anything like me they want to enter all the parameters, then try different types of simulations. I am running a coupled simulation, but I like to run the steady state flow first to get the ICs. Sometimes I need to go back and do this again.

[Michael Kennard]

If users are anything like me they want to enter all the parameters, then try different types of simulations. I am running a coupled simulation, but I like to run the steady state flow first to get the ICs. Sometimes I need to go back and do this again.

I agree. I think we could make the fields always editable but add a note in the dialog that explains when those parameters are required.