Here are excerpts from a thread of emails I have had with DHI support:
From:MikeByDHI.ca [mailto:mikebydhi.ca@dhigroup.com]
Sent:Wednesday, November 12, 2014 4:15 PM
MIKE SHE always use SZ head by layer to determine actual model calculated depths and perched zones. If depths exceed bottom of layer 1, then depth is given as bottom of layer 1.
And I would suggest using SZ head and subtract SZ head layer 1 from bottom of layer 1. Alternatively you can also use the saturation of SZ layers (I think that is included in grid output now) - which can also point to perching, if layer below top of perched layer is dry, or not fully saturated. Again, you can determine if layer 2 is fully saturated by comparing SZ layer 2 to layer 1 bottom elevation.
From:MikeByDHI.ca [mailto:mikebydhi.ca@dhigroup.com]
Sent:Thursday, November 13, 2014 11:22 AM
MIKE SHE does not calculate with “dry layers”.
All the SZ layers have some water in them. When the layer is “dry”, this default water depth is 1mm.
This methodology allows wet cells to flow laterally into ‘dry’ cells, and it allows water to drain from upper layers to lower layers.
Basically, the Phreatic Surface calculation looks for the first layer from the surface with a saturated thickness greater than 1mm.
It then creates a surface out of these elevations, or plots this value as a time series. It is not quite as simple as this, but close.
The calculation works well when the water table is in the upper layer, or when the system is dry.
However, if the water table is in a lower layer and it starts to rain, then often the saturated thickness of the upper layers increases as water builds up in the upper layers and drains after the rain stops.
This can cause confusion with the calculation of the location of the cell with the Phreatic Surface. This is the reason that the Phreatic Surface sometimes jumps up and down during the simulation. For example, the Phreatic Surface might jump several meters when irrigation is turned on, and then drop again when it is turned off.
From: MikeByDHI.ca [mailto:mikebydhi.ca@dhigroup.com]
Sent: Tuesday, December 02, 2014 9:22 AM
The result item “Depth to phreatic surface” gives the depth to the first groundwater table in the model. The problem when storing this result item is to detect whether an SZ layer is dry or not, and if dry then look at the next layer below - and continue like this until the first non-dry layer is found. The result is then the depth from the ground surface to the groundwater table in this layer. The result item could have both negative and positive values. Negative values refers to a location where the first groundwater table is below the surface, and positive values to locations where the first saturated layer is completely saturated and there are water on the surface.
In order to better understand this result item I will give you the definition of the result item and a short explanation of positive values (water on the surface). As Douglas mentioned then I would recommend you to use the actual head in a layer and subtract this with the surface to get a better understanding of the position of the groundwater table.
Definition of depth to the phreatic surface: The definition of the depth to the phreatic surface is based on a criteria for when a layer is dry. This was changed during the 2011 release, and the criteria is now:
A layer is dry if
1) The saturated thickness is less than Epsilon.
and
2) The Unsaturated thickness of the layer below is greater than Epsilon.
Epsilon is calculated as the maximum of three measures:
1) 2 x the specified Delta-H tolerance criteria for the SZ solver. The solver can't be expected to get closer to the correct solution than this value. Typically small, say, 1E-5 - 1E-3 (m).
2) A "numerical epsilon" = 1E-6 x MAX ( ABS ( all SZ layer levels in the model ) ). That's the expected precision with single-precision variables. Normally this is also small, say, 1E-5 - 1E-4 (m).
3) A partly arbitrary criteria = a fraction of the smallest SZ layer thickness anywhere in the model. This formulation makes is "less arbitrary" than a hard-coded Epsilon value. As the "fraction" we have chosen 1 % = 0.01.
For the above criteria it will normally be criteria 3 which will be used, hence 1% of the smallest SZ layer thickness.
So the above criteria defines when a layer is dry, and if the depth to the phreatic surface should be taken from a specific layer. Please note that the above calculations only refers to the result item, and it is NOT affecting the actual SZ calculations. How a criteria is fulfilled for a specific layer will determine if the depth to phreatic surface “jumps” from 1 layer to another.
Positive values of depth to phreatic surface: Positive values refers to water on the surface. This will be the case where you have upward flow from the saturated zone, or when the rainfall exceeds the infiltration capacity, and the overland flow component will not be able to move the water. Please note that a very small Manning number, or features in the topography could result in a high water column on the surface. Also look into if you have internal drain locations – drain cells flowing to internal points.
