Grid vs node time series location

MOHID Land forum. Questions and discussion related to MOHID Land
epelde.ane@gmail.com
MOHID Advanced user
Posts: 172
Joined: Thu Jan 01, 1970 00:00

Grid vs node time series location

Postby epelde.ane@gmail.com » Thu Nov 10, 2011 14:59

Hello all,

I have a doubt about grid time serie location and node time serie location.
--A grid time serie location can be created in any grid cell (regardless the cell has or hasnt a river)
--A node time serie location can be created ONLY on river network.

So, is there any ilustration or scetch about the difference between a grid time serie location in a cell on the uppest horizon and a node time serie location if in that grid cell there is a river?

And, is there any other about how is the flow calculated? I mean, mohid calculates water content for a cell and it represents the value in the center of the cell (i think i´ve read that you can choose betwenn representing in the center or in the apices). So, is this similar to finite differences method?

thanks!

Jauch
Site Admin
Posts: 98
Joined: Thu Jan 01, 1970 00:00

Re: Grid vs node time series location

Postby Jauch » Mon Nov 14, 2011 12:16

Hi Ane, :-)

--A grid time serie location can be created in any grid cell (regardless the cell has or hasnt a river)
--A node time serie location can be created ONLY on river network.


Exactly

So, is there any ilustration or scetch about the difference between a grid time serie location in a cell on the uppest horizon and a node time serie location if in that grid cell there is a river?


I don't know if there is a sketch (I need to look for it), but maybe I can clarify this question a litle bit.

A time series location file has info about cells (2D or 3D) where you want outputs.
For a 2D output (e.g. RunOff), you need to set X and Y. This can be done by the number of column and row in the grid (LOCALIZATION_I and LOCALIZATION_J), or by coordinates (COORD_X and COORD_Y).
For the 3D, the 2D position can be set the same way, but the position at the vertical axis you always need to define by the number of the layer (1 is the lower, n is the top layer). So, if you want to have the output on the top layer in a 3D geometry (porousmedia, for example), and you have 10 layers defined in your geometry, than you will set LOCALIZATION_K = 10. You need to remember that the layers are created from top to bottom. So, depending on the depth of the soil at the cell you want information, the first layer, from bottom up, that has information, can be other than 1. For example, if you set 10 layers of 10 cm each and a cell has a depth of 80 cm, than the first layer with information will be the third from bottom up (number 3).

For the nodes, you need to set the NUMBER of the NODE that you want info, not the cell where it is. You need to be careful with this.

And, is there any other about how is the flow calculated? I mean, mohid calculates water content for a cell and it represents the value in the center of the cell (i think i´ve read that you can choose betwenn representing in the center or in the apices). So, is this similar to finite differences method?


First, No. Mohid do not uses finite differences method.
Each cell is a homogenous VOLUME. So, the water content of a cell represents the volume of that cell that is occupied with water. A cell that has a theta of 0.3 says to you that 30% of the cell volume is water.
We work with fluxes too. So, the flux of water entering or exiting one cell, go through the faces of adjacent cells. To compute the fluxes we need to know the velocity of the water between two cells. The velocity is computed using the gradient of water between two cells times the conductivity. Each cell, been homogeneous, have ONE conductivity (that varies in time, when the water content varies). This conductivity is homogenous to the entire cell. But in order to have the velocity, we need to compute the conductivity on the face where the flux will pass. There are some methods available, but we usually uses the mean between the two conductivities (of each cell involved). This is a simplifyed view of the process, of course.

But the flow on the drainage network is not computed this way. It involves the manning of the channel, the slope between two consecutive nodes, the level of water on the channel, the section of each node and some other parameters, depending the method that is being used.

epelde.ane@gmail.com
MOHID Advanced user
Posts: 172
Joined: Thu Jan 01, 1970 00:00

Re: Grid vs node time series location

Postby epelde.ane@gmail.com » Wed Nov 16, 2011 09:20

Thank you eduardo! :)
So, another question about nodes. When I select a node I´m selecting a section instead of a volume (which is what we select when working with cells). So, when I make an output in one node, Im getting information of that section, right?
But, if I zoom the drainage network in the map window, I see that the down node of a reah dont match up with the up node of the following reach. Why do this happen?

ana.rosa.maretec@ist.utl.pt
MOHID Developer
Posts: 8
Joined: Thu Jan 01, 1970 00:00
Contact:

Re: Grid vs node time series location

Postby ana.rosa.maretec@ist.utl.pt » Wed Nov 16, 2011 16:44

Hi Ane,

Eduardo asked me to make a sketch about the node and reach id's. Does it help?

You're right when you say that when selecting a node, you select a cross section.

But all ID's numbers are independent. They don't have to be ordered, neither in nodes or reaches. Everything is specified in the drainage network file (*.dnt), where you can see that each reach (<BeginReach>/<EndReach> blocks) has one upstream node and one downstream node. Each node (<BeginNode>/<EndNode> blocks) can have more than 1 upstream reaches but only 1 downstream reach. The IDs don't have to be ordered.
Attachments
Picture2.png
Picture2.png (75.4 KiB) Viewed 7493 times

epelde.ane@gmail.com
MOHID Advanced user
Posts: 172
Joined: Thu Jan 01, 1970 00:00

Re: Grid vs node time series location

Postby epelde.ane@gmail.com » Wed Nov 16, 2011 17:35

Thanks for the sketch ana rosa,
Ok, I didnt know that the number of nodes and reaches didnt fit.. But, what I meant was that if you zoom a lot the network (in mohid studio map viewer window) you see rectangles which represent reaches, and their up and down stream faces dont fit with the following reach. Maybe is just because is a way of representing it in a graphical way...
And, I thought that a reach ended when the direction of the network changed or when two reaches where joined, but I see that I´m not right.
So could you tell me please, how are these formed?
Regards,
ane

ana.rosa.maretec@ist.utl.pt
MOHID Developer
Posts: 8
Joined: Thu Jan 01, 1970 00:00
Contact:

Re: Grid vs node time series location

Postby ana.rosa.maretec@ist.utl.pt » Wed Nov 16, 2011 18:49

Ok, the fit is just a graphical simplification.
Nevertheless, a reach is just a connection between 2 nodes.
Can you explain me better what you want to know? How to create a drainage network?

epelde.ane@gmail.com
MOHID Advanced user
Posts: 172
Joined: Thu Jan 01, 1970 00:00

Re: Grid vs node time series location

Postby epelde.ane@gmail.com » Thu Nov 17, 2011 08:30

Ok, so the matching matter is because of graphical simplification.
And, what I was trying to ask about the creation of nodes was, that if you look to the sketch, I undestand that the node N509 has to be there because there is a change in flow direction. And I undestand as well that a node has to be created when two reaches meet. But, N541? Because you draw a node in every cell center, but as there are reaches longer than a cell, I dont know when mohid creates a node and when dont.

And, I´m going to take advantage of this post to ask another question about the first post of eduardo. He said:
"...if you set 10 layers of 10 cm each and a cell has a depth of 80 cm, than the first layer with information will be the third from bottom up (number 3)"

Aren´t the cells´ depth the ones described as layers in the module geometry?

Thanks and regards,
Ane

Jauch
Site Admin
Posts: 98
Joined: Thu Jan 01, 1970 00:00

Re: Grid vs node time series location

Postby Jauch » Thu Nov 17, 2011 09:42

epelde.ane@gmail.com wrote:And, I´m going to take advantage of this post to ask another question about the first post of eduardo. He said:
"...if you set 10 layers of 10 cm each and a cell has a depth of 80 cm, than the first layer with information will be the third from bottom up (number 3)"

Aren´t the cells´ depth the ones described as layers in the module geometry?
Ane


Sorry Ane. I was thinking of a soil column, but wrote cell. Indeed, a cell height (or depth) is the ones described as layers in the module geometry.

ana.rosa.maretec@ist.utl.pt
MOHID Developer
Posts: 8
Joined: Thu Jan 01, 1970 00:00
Contact:

Re: Grid vs node time series location

Postby ana.rosa.maretec@ist.utl.pt » Thu Nov 17, 2011 15:39

Hi Ane,

Yes, Mohid creates a node in every cell center, starting from the specified minimum drainage area. The reaches only connect nodes.

epelde.ane@gmail.com
MOHID Advanced user
Posts: 172
Joined: Thu Jan 01, 1970 00:00

Re: Grid vs node time series location

Postby epelde.ane@gmail.com » Fri Nov 18, 2011 09:10

OK, I see.. Thank you :)

epelde.ane@gmail.com
MOHID Advanced user
Posts: 172
Joined: Thu Jan 01, 1970 00:00

Re: Grid vs node time series location

Postby epelde.ane@gmail.com » Wed Dec 21, 2011 18:15

Hello!
I´d like to create another node time serie and I wonder if the only thing I have to do is thinking which will be the node and in the nodetimeserie location file i already have, writing begin node block, node ID and endnode block. I´ve tried this and when I´ve checked the channel flow file, I still have just one column of flow data, which corresponds to the node I previously had.
Isnt it this way how I can saying all the nodes in which i want a time serie to be created?
I need a hand with this, please :)

ana.rosa.maretec@ist.utl.pt
MOHID Developer
Posts: 8
Joined: Thu Jan 01, 1970 00:00
Contact:

Re: Grid vs node time series location

Postby ana.rosa.maretec@ist.utl.pt » Wed Dec 21, 2011 19:08

Hi Ane!
For node time series, you have to put all the IDs inside the same block, i.e.:

<BeginNodeTimeSerie>
1
2
3
4
<EndNodeTimeSerie>

outputs Nodes 1,2,3 and 4.

This method is a little different from other time series because usually you must specify at least (i,j) for each location, and therefore have multiple blocks. In the DNet, only the node ID specification is needed, and they can be all inside one begin/end block.
Best regards
Rosa

epelde.ane@gmail.com
MOHID Advanced user
Posts: 172
Joined: Thu Jan 01, 1970 00:00

Re: Grid vs node time series location

Postby epelde.ane@gmail.com » Thu Dec 22, 2011 08:49

Hi rosa!
Ok, I see. Thank you very much for replying!
Regards


Return to “MOHID Land”

Who is online

Users browsing this forum: No registered users and 1 guest