A great question, which I'm sure will generate a lot of responses.   First question is to define the objectives of the model (like Sven said) and go from there picking which code you want to use.  I personally prefer MODFLOW for most projects, but here are my pros and cons.  Note I'm assuming you can use USG or NWT/SURFACT for MODFLOW not just MODFLOW-2005.  Without these codes you cannot do unstructured grid refinement or simulate drying/re-wetting of cells easily, which FEFLOW does both.

MODFLOW - Pros - easy to set up and pre/post process files; industry standard; free to use and GUIs are inexpensive to run/view/process model; modular (new packages added frequently)

Cons - currently cannot simulate complex geological features, such as angled faults and simulate steep hydraulic gradients such as rewetting/drying cells using the same code;  have to choose either USG or NWT/SURFACT capabilities.

FEFLOW - Pros - computational mesh can incorporate complex geologic features; simulate wetting/drying of cells; FEFLOW Essentials has new groundwater age feature that is similar to ZONEBUDGET

Cons - expensive (even new FEFLOW Essentials is ~$5K/license); takes longer to learn and setup; typically need more project budget/schedule to run FEFLOW models because they are more complex than MODFLOW models and take additional time

Bottom line is a groundwater modeller needs to be able to run both codes.

Well, the choice depends on the modelling task, size, scale and complexity of the modeled groundwater system, the project time frame, the future use of the model, aso. Can you give more information or some details?

I'm making a table with advantages and disadvantages of the different approaches so students can easily understand when and why use each model. I'm starting with these two because they are widely used by the research community. Would you list advantages and disadvantages for them??

A great question, which I'm sure will generate a lot of responses.   First question is to define the objectives of the model (like Sven said) and go from there picking which code you want to use.  I personally prefer MODFLOW for most projects, but here are my pros and cons.  Note I'm assuming you can use USG or NWT/SURFACT for MODFLOW not just MODFLOW-2005.  Without these codes you cannot do unstructured grid refinement or simulate drying/re-wetting of cells easily, which FEFLOW does both.

MODFLOW - Pros - easy to set up and pre/post process files; industry standard; free to use and GUIs are inexpensive to run/view/process model; modular (new packages added frequently)

Cons - currently cannot simulate complex geological features, such as angled faults and simulate steep hydraulic gradients such as rewetting/drying cells using the same code;  have to choose either USG or NWT/SURFACT capabilities.

FEFLOW - Pros - computational mesh can incorporate complex geologic features; simulate wetting/drying of cells; FEFLOW Essentials has new groundwater age feature that is similar to ZONEBUDGET

Cons - expensive (even new FEFLOW Essentials is ~$5K/license); takes longer to learn and setup; typically need more project budget/schedule to run FEFLOW models because they are more complex than MODFLOW models and take additional time

Bottom line is a groundwater modeller needs to be able to run both codes.

There is not much difference between these two codes in accuracy and generally the accuracy depends on the grid resolution. There should be no difference between the two and thus, it comes to which code the modeler has more experience with. Experience will generally reduce model generation time and the time it takes to correct mistakes, which is where the experience comes in handy.

I have been developing groundwater models since 2002 when I started with Modflow. In my next job, six years later, I was exposed to Feflow and its applications for local scale groundwater flow and solute transport modelling. I must admit that since that time Feflow is my preference, although I still use Modflow in some projects. I value Feflow for flexibility of a finite-element mesh, direct coupling with solute and heat transport and fracture modelling. Apparently Feflow is doing good job in unsaturated zone modelling, but I do not have much experience with it.

It definitely takes some time to get acquainted with concepts and applications of the Feflow platform, but I personally got benefits for the time I invested.

I am aware that Modflow is developed very fast and becomes a strong alternative for FEM, but it takes some time until new concepts are fully verified and can be used for various applications (e.g. USG and solute transport modelling).

If you have the possibility of using both, this will be an extra calibration exercise. This can be done conveniently with GIS tool (Please refer to Gossel et al., 2008). I strongly agree with all the comments of the colleagues preceding me.

Ninguno de los dos. Primero hay que saber qué Coeficientes Elásticos se pueden obtener. Si no están esos coeficientes, bien calculados, y no es fácil, los modelos dan el resultado que quiere el que lo montó.

hello

I am trying to create a 3d model using feeflow. I have data concerning the soil profil taken from in-situ testing. Does feflow create a 3d model of the subsoil using the input data that I have ?

i.e: does it predict a probable soil profil with variable soil properties or should I define each layer ?

thank you

MODFLOW is the main open source calculator for modeling groundwater flow and contaminant transport within the porous media of soil. If you are referring to the MODFLOW FORTRAN code developed by USGS, the downside is that there is no interface to be able to add the data and model from a visual standard. You do not get to see the modeling, just inputs and outputs. There are programs like Visual MODFLOW that allow to use the MODFLOW code and visualize the model process at the same time. That is what FEFLOW basically is: a similar version of Visual MODFLOW. The choice between those two would depend on the level of expertise you might want to get. I would recommend to use MODFLOW from USGS because it is open source code and there is powerful toolboxes that allow you to run simulations under different scenarios. However, if you may afford getting a license to those paid-software, perhaps having the interface may facilitate your modeling process.

The question is good and further development of both software codes will cause new comments here. According to my work experience where Feflow was mainly used the following positive points of this program could be noted:

1. Real 3d meshes to set pinching of layers, underground structures, inclined fractures.

2. Time dependence for parameters (conductivity, storage…).

3. Several possibilities to calculate free surface and unsaturated groundwater flow condition which were included in Feflow:

• setting unsaturated zone;
• setting free surface with movable or stable vertical layering with groundwater level;
• permanent option to switch model properties from saturated to unsaturated conditions.

4. Six unsaturated zone models and manually set spline model.

5. Unsaturated groundwater flow could be combined with density driven flow, heat and mass transport.

6. Rich GIS manipulation possibilities.

7. Coding for detailed engineering tasks in FEFLOW is possible through FEFLOW API functions (IFM) and allows to create and adapt additional plugins:

• -to set modes for changing simulation parameters;
• -to introduce external dependencies for transformations of boundary conditions;
• -changing values of the parameters directly;
• -program environment: C / C ++, Fortran, Python.

8. Freezing processes are available from plugin.

One BIG positive point for Modflow. If your customer is not familiar with groundwater flow preprocessor you can supply him with model environment and small GUI-interface to modify parameters. For Modflow it is possible to run model without preprocessor and license.

We did this for local authorities to forecast operations of drain network to prevent flooding of the city.

My advice is that for large and complex groundwater modelling projects, and especially when there is spatial complexity, groundwater-surface water interactions, solute transport and/or density affected flow: MODHMS wins hands down. The main advantages are the time saving and flexibility of using a regular finite difference grid, and the flexibility of processing the data (inputs and outputs) inside or outside of a Graphic User Interface (GUI).

For large and complex models, a modeller needs the flexibility to write custom code for data processing and swapping data between software packages. Being confined to a GUI is more of a hinderance than an advantage when building large and complex models.

The FORTRAN code of MODHMS (and MODFLOW family) is relatively easy to understand, and the data associated with individual model cells can be readily accessed with ad-hoc changes to the FORTRAN source code (assuming that you have permission from HGL to modify the MODHMS code).

I have used MODHMS (and MODSURFACT) in complex 3D modelling, including extremely dipping 3D strata formations; and also in highly complex 3D seawater intrusion with complex land-use patterns. It is the stability and speed of the solvers that really matters in these situations.

FEFLOW may have a more flexible mesh. There may be some situations when a finite element mesh provides a better solution, but in 10-15 years I have never needed a finite element mesh.

MODHMS has enormous and amazing capacity to provide fully integrated simulations of surface water flow, solute transport and groundwater flow. MODHMS handles unsaturated flows well, and it is numerically stable and fast in comparison to most other solvers.

Another feature of large complex models is a need for calibration. Model development and calibration may require rebuilding a model 100's or 1000's of times, and it may require pin-pointing problems in very specific cells and locations.

Automated calibration of a large model is most flexible and most powerful when managed with PEST (automated parameter estimation) and when setting up the input/output files with some help of a GUI—but ultimately these PEST jobs require very careful oversight by the modeller. PEST requires understanding of the conceptual model, the numerical problem and the existence of parameter correlation.

There are so many risks and challenges in groundwater modelling, that using finite elements seems to me to be an option that I would only consider if there were no other alternatives.

MODFLOW is a finite difference code uses gird discretization of the site and is only for flow simulations. But, FEFLOW is a finite element code uses mesh discretization system (good for heterogeneous subsurface). FEMFLOW simulates both flow and chemicals transport.

Hi researcher ,Please is there any free software of groundwater out of GMS7.1??Kindly advice me or provide me the license of GMS10.01.