NTR: 4x P01 codes for SPI Sediment Analysis [BODCNVS-2309]

[danibodc]

4 new terms requested for parameters produced from Sediment Profiling Imagery (SPI) analysis.

Details of the Sediment Profiling Imager:
https://www.oceanimagingsystems.com/docs/REMOTS-3731D-24MP.pdf
https://www.oceanimagingsystems.com/products.php#SPI

• the prism penetrates the top few centimetres of the seafloor
• obtains a vertical cross section of the sediment-water interface by imaging the sediment in profile via a water-filled prism.

At each SPI station the camera was ‘hopped’ along the sea bed at 10m intervals, providing 5 images for analysis. Images were analysed using the software ImageJ for penetration depth, apparent redox penetration depth (aRPD) and surface roughness, following the method as described in Teal et. al (2010).

Originator's Parameter names:

• Total aRPD area (cm^2)
• Average aRPD depth (cm)
• Prism penetration (cm)
• Surface roughness (cm)

These are described on p.180:
https://www.researchgate.net/profile/Jonathan-White-32/publication/269630850_Review_of_standards_and_protocols_for_seabed_habitat_mapping/links/55e06b7608ae2fac471b6de3/Review-of-standards-and-protocols-for-seabed-habitat-mapping.pdf

• prism penetration depth: gives an indication of relative sediment compaction
• sediment boundary roughness: indicates the degree of physical disturbance or biotic activity at the sediment water boundary
• sediment apparent redox potential discontinuity depth (aRPD): assesses the depth of oxygenated sediment on the bottom

Proposed P01 preferred labels:

Originator's label	Proposed P01 preferred label
Total aRPD area	Area of apparent redox potential discontinuity layer {aRPD} in the sediment by sediment profiling imager and computation using ImageJ software algorithm
Average aRPD depth	Depth (mean) of apparent redox potential discontinuity {aRPD} in the sediment by sediment profiling imager and computation using ImageJ software algorithm
Prism penetration	Penetration depth of prism in the sediment by sediment profiling imager and computation using ImageJ software algorithm
Surface roughness	Surface roughness of the sediment-water interface by sediment profiling imager and computation using ImageJ software algorithm

R.e. 'Surface roughness of the sediment-water interface by sediment profiling imager and computation using ImageJ software algorithm'

Am not sure if not specifying a matrix is ok here. The sediment-water interface is technically the floor/bed.

[danibodc]

Would require the following new elements:

• S06 - Surface roughness
  I have seen several different terms used to describe this: sediment boundary roughness, surface boundary roughness, surface roughness. 'Surface roughness' is present in the data files we have received. All appear to be synonymous.
• S06 - Penetration depth
  Similar to existing term https://vocab.nerc.ac.uk/collection/P01/current/CONEMXDP/
• S29 - apparent redox potential discontinuity (with and/or without 'layer')
• S29/S26 - sediment-water interface
  Alternatively we could use 'seafloor' which is proposed as a new S18 here: Legacy Sea-floor depth parameter codes to be mapped to their atomic components (BODCNVS-1860) #213, however we might need to be broader e.g. 'bed'. However the literature related to SPI favors 'sediment-water interface'.
• S04 - sediment profiling imager
• S05 - computation using ImageJ software algorithm

[gwemon]

Thanks @danibodc. A few comments and questions in relation to this ticket

"Penetration depth" is a useful concept to have in S06. However as a stand alone term, it needs to be defined carefully because it can be applied to multiple domains and take slightly different meaning when associated with an object of interest and matrix in the P01 combination. QUDT does not have a broad "Penetration depth" in its collection of quantity kinds but it has a more specific term "https://qudt.org/vocab/quantitykind/LondonPenetrationDepth"

Regarding the associated P01, I would prefer e.g.: "Penetration depth of the instrument prism in the sediment by etc" - but do we even need to specify "prism"? also do we need to specify by sediment profiling imager and computation using ImageJ software algorithm? A more generic alternative could be: "Penetration depth of the instrument in the sediment"?

I note though that you quoted that "prism penetration depth: gives an indication of relative sediment compaction"

[gwemon]

For:
Area of apparent redox potential discontinuity layer {aRPD} in the sediment by sediment profiling imager and computation using ImageJ software algorithm
and
Depth (mean) of apparent redox potential discontinuity {aRPD} in the sediment by sediment profiling imager and computation using ImageJ software algorithm

I think this works well. It could be that we need to use the S06 for Depth (spatial coordinate) here? Because it is the depth at which something is rather than the depth of something? Not sure...

[gwemon]

For: Surface roughness of the sediment-water interface by sediment profiling imager and computation using ImageJ software algorithm

Yes, I think that works well. Surface roughness would be a useful re-usable concept to have in S06 although it might be tricky to define? I see that its expressed as a length but could not find a definition that relates it to a length in my (very quick) search.

for the S29 yes I do like sediment-water interface. Again this is a useful re-usable concept to have. I agree that it could also be seafloor or sediment (bed seems strange in tis context without a "the" in front of it. But sediment-water interface would get my vote.

[alicecbodc]

Thanks everyone for taking a look at these codes! They look great. I've got a couple notes for context:

• Surface roughness calculation - from the research I have done into ImageJ, what I believe has happened is that the originator has used a plugin to effectively turn the interface into a line graph in cross-section (often called rugosity or Ra in geomorphology). However, I found the following link to a study of rugosity by SPI analysis, which calculated surface roughness (here, "Boundary roughness - maximum minus minimum penetration depth for each replicate image") https://apps.ecology.wa.gov/publications/documents/0703025.pdf, page 28. This isn't quite as accurate as rugosity as a calculation, so it could very depending on the time the originator spent analysing the images and modelling that sediment surface as a line. Ra would be the gold-standard of roughness analysis, whereas the penetration depth analysis outlined above would be a lot less accurate.

• This link here goes into a little more detail on Ra (surface roughness): https://get-it-made.co.uk/resources/surface-roughness-explained#:~:text=It%20is%20calculated%20as%20the,a%20surface's%20peaks%20and%20valleys. - this isn't scientific, however it draws a lot of parallels to some available ImageJ plugins for surface roughness in the way that it is calculated as a length. Without going back to the originator, I'm not sure we would know exactly what algorithm is being used here

• Sediment-water interface vs. other terms - with my geology hat on, I would agree that sediment-water interface encompasses this more comprehensively than "bed", "seafloor" or "sediment" because of the transitory nature of the interface

I hope this helps - do let me know if you need any more info on the above!

[danibodc]

Hi @gwemon and @alicecbodc, just coming back to this, Jak has kindly agreed to pick this up for me.

Will go ahead with the S29 sediment-water interface, and agree about using Depth (spatial coordinate). We would still need to use S07 'mean', so the P01 might look strange with two things in brackets - Depth (spatial coordinate) (mean)... , but is accurate.

I also like the more generic proposal for 'Penetration depth of instrument in the sediment' - this could be widely re-used for various corers e.g. gravity corers etc.

Proposed P01 pref labels:

Total aRPD area:
Area of apparent redox potential discontinuity layer {aRPD} in the sediment by sediment profiling imager and computation using ImageJ software algorithm

Average aRPD depth:
Depth (spatial coordinate) (mean) of apparent redox potential discontinuity layer {aRPD} in the sediment by sediment profiling imager and computation using ImageJ software algorithm

Prism penetration:
Penetration depth of instrument in the sediment

Surface roughness:
Surface roughness of the sediment-water interface by sediment profiling imager and computation using ImageJ software algorithm

[gwemon]

@danibodc it looks good to go I think. In the new models, the S07 is not in parenthesis so it won't look so weird :-)
I poundered if we really needed to embed aRPD in the preferred label but I guess it is a useful short acronym to have there.

[jakjonbodc]

Suggested new term definitions for discussion:

S06 surface roughness - The measurement of the smoothness of a surface's profile. - Done

S06 Penetration depth - The distance to which an object, force, or electromagnetic radiation travels into a material or medium. - Done

S29 apparent redox potential discontinuity layer (codalt: aRPD) - A layer of sediment bounded by the sea floor and a rapid change in redox potential, typically marked by a noticeable color transition which indicates the boundary between oxygenated and oxygen-depleted conditions in the sediment. - Done

An oxygen-rich layer of sediment bounded by the sea floor and the interface between oxygenated and oxygen-depleted material, which is typically marked by a noticeable color transition.

S26 sediment-water interface - The boundary of the seabed sediment with the body of water.

S04 sediment profiling imager - I think that a suitable S04 already exists, S041035.

S05 computation using ImageJ software algorithm - The parameter has been calculated using the ImageJ software algorithm - Done

[jakjonbodc]

note - the Redox Potential Depth (RPD) is termed the apparent Redox Potential Depth (aRPD) when estimated visually and not determined with the use of redox measurements.

[gwemon]

@jakjonbodc Adding for reference useful paper shared by @ahpanton in Teams: https://www.sciencedirect.com/science/article/pii/S1001627915600087

[jakjonbodc]

Revised descriptions:

S29 - apparent redox potential discontinuity (aRPD) - A transition of sediment colour in the seabed used as an indicator of a change in dissolved oxygen content from oxidized surface sediment to reduced deeper sediment. - Done

https://core.ac.uk/download/pdf/250026502.pdf - aRPD appears to be defined as the space between the sediment-water interface and the transition in colour of sediment. Dropped 'layer' as this term does not appear to be used alongside aRPD in literature.

S05 - computation using ImageJ software - The parameter has been calculated using ImageJ, an image processing and analysis software. - Done

[jakjonbodc]

S29 - PE001562 - apparent redox potential discontinuity - A transition of sediment colour in the seabed used as an indicator of a change in dissolved oxygen content from oxidized surface sediment to reduced deeper sediment.

S18 - S1800139 - apparent redox potential discontinuity - A transition of sediment colour in the seabed used as an indicator of a change in dissolved oxygen content from oxidized surface sediment to reduced deeper sediment.

S05 - S050269 - computation using ImageJ software - The parameter has been calculated using ImageJ, an image processing and analysis software.

S06 - S0600339 - Surface roughness - The measurement of the smoothness of a surface's profile.

S06 - S0600340 - Penetration depth - The distance to which an object, force, or electromagnetic radiation travels into a material or medium.

[jakjonbodc]

S26 - MAT02001 - sediment-water interface - The boundary of the seabed sediment with the body of water.

[jakjonbodc]

'sediment-water interface' posted to wrong vocabulary, should be an S29 term.
Description revised to 'The boundary between the sediment layer and the overlying water in a water body.' as it does not have to be specific to seabed.

[jakjonbodc]

New term created:
S18 - S1800140 - sediment-water interface - The boundary between the sediment layer and the overlying water in a water body.
S29 - PE001563 - sediment-water interface - The boundary between the sediment layer and the overlying water in a water body.

[jakjonbodc]

parameter.definition suggestions:

Total aRPD area - A sediment profile imagery parameter defined as the area enclosed by the apparent redox potential discontinuity (aRPD) and the sediment-water interface.

Average aRPD depth - A sediment profile imagery parameter defined as the depth of the apparent redox potential discontinuity (aRPD) from the sediment-water interface.

Prism penetration - unavailable - I think that this would basically be a rewording of the full_title and not add value to the vocab term.

Surface roughness - A quantification of the deviations in the direction of the normal vector from the sediment surface.

also considering: A quantification of the irregularities of the sediment surface geometry.

[gwemon]

@jakjonbodc could you point me to the reference source you used to define Total aRPD area? I could not find it in the links provided in this ticket but maybe I missed it?

[gwemon]

For Average aRPD depth -
which we are modelling as "Depth (spatial coordinate) (mean) of apparent redox potential discontinuity layer {aRPD} in the sediment by sediment profiling imager and computation using ImageJ software algorithm"

I'd suggest the following definition based on reference in the linked documnet in the original request (Curtis and Coggan, 2007, Sediment Profile Imagery in Review of standards and protocols for seabed habitat mapping):

Proposed definition: A Sediment Profile Imaging (SPI) parameter derived from image analysis using the Image J software, defined as the average vertical distance below the sediment-water interface, at which a discontinuity of colour is observed, providing an assessment of the average thickness of the oxygenated sediment surface layer.

[gwemon]

Prism penetration - which we are modelling as the more generic: "Penetration depth of instrument in the sediment"

Could be defined as: "The maximum vertical extent to which an instrument penetrates in the sediment. In Sediment Profile Imaging data, this is sometimes refer to as the prism penetration depth which gives an indication of relative sediment compaction."

[gwemon]

Surface roughness which we modelled as: "Surface roughness of the sediment-water interface by sediment profiling imager and computation using ImageJ software algorithm"

Could be defined as: "A quantification of the irregularities in the geometry of the sediment surface, determined using the ImageJ software package. It is used as an indicator of the degree of physical disturbance or biotic activity at the sediment water boundary."

[jakjonbodc]

@gwemon - area of aRPD shown in section 3.8 of https://core.ac.uk/download/pdf/250026502.pdf. I will add to this description 'derived from image analysis using the Image J software' as you suggested above for the 'depth of aRPD' parameter.

[jakjonbodc]

Average aRPD depth - thankyou for the feedback I will proceed with the definition you have suggested and will add detail to the 'area of aPRD' parameter as described above.

[gwemon]

Thanks @jakjonbodc for the reference. From looking at the document I understand that "Total aRPD area" requested here is what is being measured by image analysis in order to obtain the mean depth of the aRPD, by dividing the observed area by the width of the analysed image (see page 11 of the publication: "The mean depth (area/image width) of the reworked oxidized layer defined by the colour shift from brown/yellow to grey/black. Here it is important that the underlying sediment layer not is contrast to yellow. Just an oxidized thin layer of a spring bloom and/or fecal pellets will just score for fecal (se above) not as measured RPD > 0.1 cm, independent of the thickness of such a layer."
Is that your understanding too?
If so then this is more a metadata parameter than an environmental parameter... It is a measurement made on an image of a given width that is then used to determined an environmental variable (the mean depth of the aRPD).

If this is the case then we need to review the P01 preferred label.

[gwemon]

In the reference document, the area is effectively defined as "area of the reworked oxidized layer defined by the colour shift from brown/yellow to grey back"

I don't think that Area of aRPD layer is correct because aRPD is a discontinuity (line and maybe sometimes layer) but not what is being measured here which is the area of the oxidized layer (above the aRPD) over the width of the image.

[jakjonbodc]

Hi Gwen, Thanks for the feedback. Based on the dataset this request has come from I think that average depth was determined in the way described in https://core.ac.uk/download/pdf/250026502.pdf, i.e. mean depth = area/ image width. Given that the aRPD is a discontinuity line/ layer I think that the preflabel should be reviewed, as the area measured appears to be that of the oxidized layer based on the aRPD by image analysis, not the aRPD itself.

[gwemon]

For the area P01 I would suggest:

Area of oxidized layer [in the digital image] by sediment profiling imager and computation using ImageJ software algorithm

A sediment profile imaging parameter defined as the area enclosed between the apparent Redox Potential Discontinuity (aRPD) and the sediment-water interface determined by the ImageJ software over the width of the image being analysed. This parameter is used to calculate the mean vertical depth of the aRPD.

I hesitate about adding "digital image" to S21 and S26 because it would need a review of the definitions of these vocabularies which we might not be able to do right now. So for the time being I am proposing to have a P01 of the form:

Area of oxidized layer by sediment profiling imager and computation using ImageJ software algorithm

Please also associate this parameter to this P02 group: https://vocab.nerc.ac.uk/collection/P02/current/TMES/

[jakjonbodc]

Thanks for this Gwen, I'm happy to proceed with the P01

I think that given the S05 and S06 in this parameter we could safely leave out 'digital image', as the area being taken from an image of the sediment is implicit in the given method, unless you can think of any issues with this?

Further to that, I think it is important that we capture that this describes the oxidized layer as defined by the position of the aRPD, as I mentioned there does seem to be questions in the community about the reliability of the aRPD as an indicator of DO content compared to other methods. I think determination of area based on position of the aRPD is again implied by the S05 & S06 and is detailed in the definition.