Page 1 of 1

Using Hydrus 2D to Output Vertical Velocity Profile

Posted: Tue Jun 05, 2018 8:22 pm
by MichaelF

I am working on a 2D simulation of a cross section of a micro-catchment (200 cm across x 35 cm deep) using a time varying varying head boundary to simulate the measured water level within the catchment. I am interested in predicting the infiltration rate of a complex catchment with 5 separate soil layers (see img) during different water levels during the simulation. To accomplish this, my logic was to create a Velocity profile across the boundary of the catchment, which could be multiplied by surface area of a micro-catchment to calculate the discharge at any given timestep. This velocity profile could approximate flow out of catchments of varied geometry given an accurate stage to volume of each catchment. I am converting the boundary.out file for my atmospheric boundary that also simulates the water level when h is positive to this velocity profile, but had some questions about the output file.

The boundary.out file has a Q value with the listed units of [V/T]. I am unclear what this Q value is used for, and why is does not have units of [L^2/T]? It also can not calculate this Q values by multiplying the velocity by the node length.

I have been using the v [L/T] columns for my boundary to create the vertical velocity profile. I wanted to make sure that the output velocity of the water would not need to be multiplied by the volumetric water content in order to calculate a flux.

My initial simulations under-predicted the total infiltration rate out of 7 microcatchments by 25-60%. I am also curious if there is any documentation on translation from 2D simulations to 3D application where radial flow from the catchment comes into play.

Thanks for reading my lengthy question! I'm happy to provide any additional details.

Re: Using Hydrus 2D to Output Vertical Velocity Profile

Posted: Tue Jun 05, 2018 9:27 pm
by Jirka
Q has units of [V/T] where V is L2 for 2D problems and L3 for axisymetrical problem. This is explained at the beginning of the file (I believe) and certainly in the Technical Manual. This is the value of a nodal flux (which is what the FE method recognizes). To get boundary velocity, you need to divide this by the boundary length associated with the node (which you can find in

Note that you can use Meshlines to get fluxes through any line within the transport domain, as well as across any part of the boundary.

I do not see how you could convert your 2D domain into radially symmetric domain without significant simplification. Note that you need an axis of symmetry.


Re: Using Hydrus 2D to Output Vertical Velocity Profile

Posted: Tue Jun 05, 2018 10:22 pm
by MichaelF
Thank you for the prompt reply. I had missed that the file included a width between each node. It seems like the boundary velocity should averaged by height intervals should suite my needs to predict discharge from a complex basin.

I agree that my model is overly complicated to simulate as asymmetric in its current form, but i will try to simplify to see how much the velocity as a function of depth changes in a radially symmetric domain. I was curious if you had encountered others attempting to use a 2D model to create a velocity profile that can be applied to more complex 3D shapes? I haven't found any information in the the literature so far, but thought it might be a useful application of 2D models.

Thanks again for the help!