Assignment or adjustment of DEPTH in EDIT section not working

[EdmundStephens]

[v2025.10] Assignment or adjustment of DEPTH in the EDIT section is not working. This is a useful way of introducing dip or gravity potential differences (especially since ADDZCORN is still not available). Formally speaking the DEPTH array represents the potential of a grid cell, by default this is the centroid of the corner points, but may differ: i) in upscaling considering weighting on net sands, ii) in order to represent a dipping structure within a simple 'simbox' grid, iii) models of centrifuge experiments etc. where there is a different (and variable) gravity potential.

Test model.

TEST_DEPTH-F.zip

At the moment methods for altering the DEPTH array result in errors or do not have any effect.

EQUALS results in an error

EQUALS
	DEPTH		8900		3  4	1  8	1  8	/
/

Error: Problem with keyword EQUALS
In TEST_DEPTH-F.data line 205
Target array DEPTH is not supported in the EQUALS operation

Error: Unrecoverable errors while loading input:
Problem with keyword EQUALS
In TEST_DEPTH-F.data line 205
Target array DEPTH is not supported in the EQUALS operation

ADD results in an error

ADD
	DEPTH		-100			3  4	1  8	1  8	/
/

Error: Problem with keyword ADD
In TEST_DEPTH-F.data line 205
Target array DEPTH is not supported in the ADD operation

Error: Unrecoverable errors while loading input:
Problem with keyword ADD
In TEST_DEPTH-F.data line 205
Target array DEPTH is not supported in the ADD operation

OPERATE results in a different type of error

OPERATE
	DEPTH		3  4	1  8	1  8		ADDX		DEPTH		-100		/
/

Error:
An error occurred while creating the reservoir properties
Internal error: INFO: 'DEPTH' is not a double precision property.

Error: Unrecoverable errors while loading input:
INFO: 'DEPTH' is not a double precision property.

OPERATER is allowed but does not actually update the DEPTH array

OPERATER
	DEPTH		2	ADDX		DEPTH		-100		/
/

This is the resulting DEPTH array from Flow:

This is what the DEPTH array should look like (same colour scale):

[bska]

The fundamental reason for this behaviour is that the simulator does not maintain a DEPTH array. Instead it only has the cells' centre point depths and derives the DEPTH output array from those. Adding a separate DEPTH array/property would require a bit of work and while I'm not saying that we won't do that, we've not seen a model that needs it before you opened this issue.

[EdmundStephens]

DEPTH is independent of the corner points centroid in a number of situations:

• Simulations of core plug centrifuge experiments, DEPTH is assigned as the rotational potential varying with radius, which looks like varying gravity across the plug (e.g. along z axis)
• In larger upscaling situations, such as basin scale models, DEPTH should be the centroid of the porosity, which is not necessarily the same as that of the corner points
• Making a 'simbox' type grid, where dip is created by assignment to the DEPTH array ... such models are sometimes desirable if these enable automated processes, such as ensemble / regression exercises involving variation of structure flexes or fault displacements
• As academic tool to 'switch off' or otherwise vary the gravity force (e.g. giving all cells the same DEPTH removes gravity)
• Synthetic dual porosity where offset 'grids' are linked by NNC connections setting e.g. TRAN_MF = SIGMAV * PERM * CELLVOLUME, TRAN_TH = SIGMATH * THCCON * CELLVOLUME. These types of model enable Flow to model dual porosity type models; the same method can be extended to a stack of grids to allow TRPLPORO, QUADPORO and NMATRIX type models to be constructed

The DEPTH array by default is equal to the centroid of the corner points but this is a bit of a misnomer for the 'cell hydraulic potential'.

[totto82]

Note that DEPTH was just implemented. OPM/opm-common#5181. And it will be available in the next release.