Why so many different properties in S06 related to concentrations/proportions/ratios etc? (BODCNVS-1791) #74
Comments
|
related to this, when I search 'ratio' under chemical parameters builder it is not found as a concept. So there is a relationship between concentration and ratio too, under proportion? btw I suspect soil scientists are very used to saying "ratio of" and would find it strange to use 'concentration of' (or 'proportion of'). So I think ideally it would be a parameter if appropriate. Here are some commonly used soil science examples -
|
|
some of our methods state "Report elemental concentrations (% or mg/kg of element) on an oven-dry basis" - which makes me wonder when if ever in the case of soil chem we would use 'proportion of'. People are very familiar with 'percentage of' , thought I understand that is the UoM. If we said "concentration of...." the related UoM is percentage. |
danibodc
changed the title
|
@dr-shorthair @meganrwong Following your request, we had a review of our recent addition of "Mass concentration" to the S06 vocabulary and of the P01 terms which are associated with it (all related to the eREEFs model output) and we realised that we introduced an inconsistency in the way we model this kinds of properties. Since the variables were all related to particles we should have modelled them as "Mass of particles xyz per unit volume of sediment by model prediction" and not used "Mass concentration". We will correct this. |
|
To avoid confusion we are going to deprecate "Mass concentration" since it is not used as part of our PUV semantic model and we will redirect users to the QUDT concept "Mass concentration" whenever such a concept is needed (e.g. when using I-ADOPT modelling) instead of duplicating it. |
|
Your question about proportion vs concentration is a good one and not straightforward to answer! I seem to recollect, that Roy had considered in the past when we started remodeling legacy terms whether we should call them all "Concentration". We opted to keep the two terms and make the Proportion parameters more explicit by adding Proportion by dry weight, Proportion by volume, Proportion by area etc.... It would be interested to know what the different communities' preferences or views would be and how we could better harmonise. |
|
Much thanks for this Gwen, So this is where we are at atm with using 'concentration' v 'proportion', in light of your comments, with some examples -
That is where our list is at atm anyway, feedback welcome :) |
|
To me the difference between concentration and proportion is that the dimensions of the quotient and divisor for concentration are different e.g. mass/lengthlengthlength, but the dimensions of the quotient and divisor for proportion are the same e.g. mass/mass. As an aside, one of my errors that I still regret was not to differentiate mass concentration, mole concentration and volume concentration in the way CF have to keep the mapping between term and dimensionality 1:1. This was a hangover of another error - enforcing a 1:1 mapping between P01 code and units - that I inherited and was partially fixed by restricting the practice to within BODC. |
|
regarding @meganrwong last message.... Following the dimensionality principle, I would agree with concentration for cmol/kg, cmol/kg etc. but would use proportion for mg/kg. Note that geology is more concerned with solid chemistry and so reporting in terms of mass/mass is much more common than in oceanography where reporting in terms of mass or moles/kg is more normal. |
|
Looking further into this now that you mentioned the units/dimension association Roy, I have checked what we have at the moment and all our "Proportion xyz" codes have units expressed as "percent". It could be that this was a consequence of a decision taken when we remodeled the terms. Here ae the units listed against parameter codes where Property='Concentration" as it stands at the moment: |
|
And this is the list of all "Concentration" concepts that are mapped to units of "Percent": |
|
I think this may possibly be improved with no consequence for legacy data by changing concentratiom to proportion where the dimensions are unquestionably mass/mass. The only tripwire to watch out for is where the 'concentration' could be moles/mass rather than mass/mass in the legacy data. There are times when I wish I'd been exposed to Jonathon Gregory's rigorous physics-based approach used in CF Standard Names in 2004 when I was formulating the P01 semantic model rather than in 2008 when I actually encountered it! |
|
As you may have gathered, a can of worms is being opened here....... |
|
The no-brainers that could be changed are things like 'Concentration of lipids per unit wet weight of biota {Tursiops truncatus (ITIS: 180426: WoRMS 137111) [Sex: female Subcomponent: muscle tissue]}' that could in no way be expressed as moles rather than mass. The ones that make me cringe a bit are things like 'Concentration of total iron {total_Fe CAS 7439-89-6} per unit dry weight of sediment'. In all my experience of this measurement it would be reported in units of per cent or ppm pertaining to mass of iron per mass of sediment. within the rules of the game, some bright spark could report this measurement in moles/mass. However, I'm starting to think that if they did, it would be no worse than our having a mixture of concentrations in mass/volume and moles/volume and that having 'proportion' rather than 'concentration' might provide helpful guidance and provide damage limitation for past errors. It would at least lead towards a simple definition of 'concentration'!!! |
|
All sorts of thoughts are starting to fly around my head here... If we have chlorophyll in ug/l and chlorophyll in ug/kg should one be a concentration and the other a proportion??? I tried to steer P01 in the direction of separating 'concentration per unit volume' and 'concentration per unit mass' to deal with this. situation but I'm now driving myself into a corner where 'amount per unit mass' should be a proportion for chlorophyll, but a concentration for metals or nutrients that are moles/kg????? It all comes down to P01 having more ambiguous terminology than CF..... I'm thinking that creating a mapping between CF terminology, dimensionality and P01 terminology for concentrations/proportions/abundances might provide insight into the issue being discussed here which could provide a way forward that would improve interoperability between P01 and CF Standard Names. We can only fix things without making matters worse by having a clear understanding of the problem and producing a clear set of rules for current usage that don't screw up legacy...... |
|
Our documentary base mostly uses the term 'ratio' in the labels. |
|
Realised that defining proportion versus concentration in terms of dimension was a red herring based on my trying to remember thought processes of 20 years ago contaminated by working on CF Standard Names. So don't panic Gwen! I've been trying to remember the history.... Concentration started out as a 'no-brainer' for amounts of a chemical substance per quantity of matrix Then came GeoSeas and the need to cover things like classical grainsize where mass is a proxy for a count of occurrences of a physical property. So proportion was introduced as using concentration in the context of grainsize is semantically ridiculous. We then met radioactive substance data where the measurement is the number of decays per unit time and so 'activity' was introduced as this is significantly different to concentration (concentration for a given isotope activity depends on half life) I see a ratio as the division of one concentration/proportion/activity by another. My advice to @meganrwong to use 'proportion' for what are obviously chemical measurements was from my thinking along the wrong lines and totally incorrect! Getting clear definitions of these terms would be useful as would increasing awareness of what CF are doing (e.g. my suggested mapping) so any future changes bring the vocabularies together rather than their drifting further apart. |
|
(Shouldn't you be in bed @roy-lowry ?) |
|
Here is a slightly different example where some judgement was required - just to point out the challenge of applying a 'hard and fast' rule re when 'concentration' would be used.
Although all are reported in molc/m3, to me it would not read right to have 'concentration' of for no.3. I think it would just be 'Residual quick neutralizing capacity' (S29) 'on a moist soil basis' (S03 - sample preparation) |
|
@meganrwong Now you are providing the reason why there are so many S06 entries ;-) Your last example is very similar to the parameter called 'total alkalinity' in oceanography, which is a measure of a water sample's ability to neutralise acid. This has its own S06 entry (S0600283). At one time this was reported in units of microequivalents/litre, but now concentration units (umol/kg) are used. However, as you say, including concentration it a description of alkalinity would not make good reading. |
We are attempting to define some soil chemical parameters using the NVS Parameter Usage model.
Most soil chemical parameters are of the nature 'concentration or abundance of X using method Y on sample prepared by Z'.
We don't expect to find all the required elements for soils in the NVS S03 S04 S05 S25 S26 vocabs, but S02 S06 S27 should service us.
However, S06 has a lot of terms related to proportions, concentrations, and ratios.
It is not clear to us which of these base properties to use, when, to build the parameters.
For example, what is the relationship between
or between
Are these all somehow in a specialization hierarchy, perhaps with
at the top?
Is there an general rule to explain what parameter is a 'concentration' and what is a 'proportion'?
@meganrwong @abhritchie @ljgregory are involved.
The text was updated successfully, but these errors were encountered: