LPF package convertible upstream option

[Laura Rod]

Hello

My model is not running when I set the 1st layer as convertible (Unconfined aquifer). However, when I tested the Convertible Upstream option in the LPF package my model finally ran (still didn't converge tough). In what cases do you recommend to use the Convertible Upstream option from the LPF package? I'd like to know if it is recommended for a 2-layer unconfined aquifer before changing other parameters in my model to make it converge.

[HelenaH]

Hi Laura,

If it's just a simple 2-layer model, you should be able to use convertible layers, even before changing your parameters. However, I've experienced issues with using convertible layers if the layers are too thin, and/or if there are dry cells. Though I've only had this issue with ugrid (unstructured grid) models it might also occur using a structured grid. 

All in all, there should be no issues using the LPF package and convertible layers, however there might be other issues with the model resulting in MODFLOW ending before running the model. 

[Laura Rod]

On 9/17/2022 at 3:52 AM, HelenaH said:

Hi Laura,

If it's just a simple 2-layer model, you should be able to use convertible layers, even before changing your parameters. However, I've experienced issues with using convertible layers if the layers are too thin, and/or if there are dry cells. Though I've only had this issue with ugrid (unstructured grid) models it might also occur using a structured grid. 

All in all, there should be no issues using the LPF package and convertible layers, however there might be other issues with the model resulting in MODFLOW ending before running the model. 

Thanks HelenaH. My model has a Ugrid as well. I ran the model with convertible negative and now I'm running it again with convertible layer and it takes longer, the convertible upstream option is much faster. I am waiting for the results of this second run to see if it makes any difference on my heads.

[Laura Rod]

Thanks Helena

I had to modify my model to include a third layer (Ugrid) to avoid some starting head issues and changed the parameters as follows: Layer 1 convertible negative, Layer 2 convertible upstream, and Layer 3 confined. Not sure if I'm right, but what I understand is that the convertible upstream option just calculates the transmissivity assuming that the saturated thickness is the same, while the convertible negative recalculates the transmissivity with a different saturated thickness depending on the water table height, which would be ideal for the first layer of my unconfined aquifer. However, the computational time is huge! let's see what I get from this run test.

Unfortunately, the LPF package documentation is not very clear about the differences between the three convertible options.