design formative element dictionary content, goals, editing suggestions

[dylanbeaudette]

Sub-Issues:

• possibly normalize dictionaries such that each term is tied to a master "root" (this is complicated because the same roots can be used in slightly different ways across levels of ST)
• add "root" column to existing dictionaries for iteration vs. assumption that "connotation" is replicated

[brownag]

So, I am proposing we remove the redundancy from the formative element table. We already are partitioning meaning by the levels of ST... but indeed this is only the tip of the ice berg.

I don't know that we want to go the route of doing this for all the instances this kind of thing affects. Here is an example.

Currently we have:

sulfo,sulfur,presence of sulfides or their oxidation products,, 
sulfi,sulfur,presence of sulfides or their oxidation products,, 
sulf,sulfur,presence of sulfides or their oxidation products,,

Which produces identical results to the simplest case:

sulf,sulfur,presence of sulfides or their oxidation products,,

Now, sulf- is an example of a formative element whose meaning varies depening on the great group it is used with. There are some patterns, but they are not perfect.

sulfo- always means presence of a sulfuric horizon

sulfi- always means presence of sulfidic materials

sulfu- always means presence of a sulfuric horizon (only one instance -- sulfudepts)

sulfa- varies depending on the soil order (entisols cannot have a sulfuric horizon)

• sulfaquerts: presence of a sulfuric horizon or sulfidic materials,,
• sulfaquepts: presence of a sulfuric horizon
• sulfaquents: presence of sulfidic materials

As it stands we have a mixture of degrees of specificity of these definitions... some refer to actual existing definitions in soil taxonomy, whereas others simplify the meaning. "presence of sulfides or their oxidation products" is not wrong, with sulfur as the primary connotation, but saying either of the diagnostic materials/horizons is present will be wrong for some subset if we limit to just sulf-.

We can go and define all of instances I gave above as their own definition, and it works, or we can go the route of not actually identifying the specific relevant definitions and criteria in these more generalized explanations. I know @dylanbeaudette has discussed different levels of detail for this output, and I think the current defnitions we are using probably span what would be a couple different levels as is.

[brownag]

From Soil Taxonomy (2003), CHAPTER 6 The Categories of Soil Taxonomy, p. 119. https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/nrcs142p2_051232.pdf

I think we need to work these definitions into the ST_formative_elements metadata to explain "intragrade" "intergrade" "extragrade" logical columns. Then we should ensure that the records are assigned as "correctly" as possible.

Typic subgroups.—These are not necessarily the most extensive subgroups, nor do they necessarily represent the central concept of the great group. In some taxa typic subgroups simply represent the soils that do not have the characteristics defined for the other subgroups.

Intergrades or transitional forms to other orders, suborders, or great groups.—The properties may be the result of processes that cause one kind of soil to develop from or toward another kind of soil or otherwise to have intermediate properties between those of two or three great groups. The properties used to define the intergrades may be:

1. Horizons in addition to those definitive of the great group, including an argillic horizon that underlies a spodic horizon and a buried horizon, such as a thick layer of organic materials that is buried by a thin mineral soil; or
2. Intermittent horizons, such as those described in the section of chapter 1 that deals with the pedon; or
3. Properties of one or more other great groups that are expressed in part of the soils but are subordinate to the properties of the great group of which the subgroup is a member. One example of different depths of saturation and reduction was given earlier. Another example might be an Alfisol that has an ochric epipedon a little too thin or a little too light in color to be a mollic epipedon. This feature could result from an invasion of grassland by forest or the reverse, from the coexistence of both grass and forest, or from the erosion caused by human activities.

Extragrades.—These subgroups have some properties that are not representative of the great group but that do not indicate transitions to any other known kind of soil. One example of an overthickened mollic epipedon was given earlier. Other examples are soils that are very shallow over rock (Lithic) or soils that have high amounts of organic carbon (Humic).

[brownag]

I think some of the explanations need revision.

For instance "typic" probably should read "central theme of great group concept" (not "subgroup").

library(SoilTaxonomy)
cat(explainST('typic endoaqualfs'))
#> typic endoaqualfs
#> |     |   |  |                                                                                      
#> central theme of subgroup concept                                                                   
#>       |   |  |                                                                                      
#>       ground water table                                                                            
#>           |  |                                                                                      
#>           characteristics associated with wetness                                                   
#>              |                                                                                      
#>              soils with an argillic, kandic, or natric horizon

Also, we should consider using/replacing some the connotations included in latest versions of the Keys

e.g. "Hapl" -> "minimum horizon development" rather than "central theme of subgroup concept"

Also consider adding historical formative elements no longer used in modern versions of Taxonomy

[dylanbeaudette]

Add 'wassi' formative element.

[brownag]

Add 'wassi' formative element.

"wassi" is not a formative element. "wass" is the formative element, followed by "i" (h_i_stosols) or "e" (_e_ntisols).
It is already in the suborder dictionary.

SoilTaxonomy/misc/formative-elements/dictionaries/suborder.csv

Line 31 in 6c095c4

wass,water,daily under shallow water,,

library(SoilTaxonomy)
data("ST_unique_list")
lapply(ST_unique_list, \(x) x[grep("wass", x)])
#> $order
#> character(0)
#> 
#> $suborder
#> [1] "wassists" "wassents"
#> 
#> $greatgroup
#> [1] "frasiwassists"  "sulfiwassists"  "haplowassists"  "frasiwassents" 
#> [5] "psammowassents" "sulfiwassents"  "hydrowassents"  "fluviwassents" 
#> [9] "haplowassents" 
#> 
#> $subgroup
#>  [1] "fibric frasiwassists"        "sapric frasiwassists"       
#>  [3] "typic frasiwassists"         "fibric sulfiwassists"       
#>  [5] "sapric sulfiwassists"        "typic sulfiwassists"        
#>  [7] "sulfic haplowassists"        "fibric haplowassists"       
#>  [9] "sapric haplowassists"        "typic haplowassists"        
#> [11] "lithic frasiwassents"        "psammentic frasiwassents"   
#> [13] "thapto-histic frasiwassents" "fluventic frasiwassents"    
#> [15] "aeric frasiwassents"         "typic frasiwassents"        
#> [17] "sulfic psammowassents"       "lithic psammowassents"      
#> [19] "fluventic psammowassents"    "aeric psammowassents"       
#> [21] "typic psammowassents"        "lithic sulfiwassents"       
#> [23] "haplic sulfiwassents"        "thapto-histic sulfiwassents"
#> [25] "fluventic sulfiwassents"     "aeric sulfiwassents"        
#> [27] "typic sulfiwassents"         "sulfic hydrowassents"       
#> [29] "grossic hydrowassents"       "lithic hydrowassents"       
#> [31] "thapto-histic hydrowassents" "typic hydrowassents"        
#> [33] "sulfic fluviwassents"        "lithic fluviwassents"       
#> [35] "thapto-histic fluviwassents" "aeric fluviwassents"        
#> [37] "typic fluviwassents"         "sulfic haplowassents"       
#> [39] "lithic haplowassents"        "aeric haplowassents"        
#> [41] "typic haplowassents"

cat(explainST("sulfic haplowassents"))
#> sulfic haplowassents
#> |      |    |   |                                                                                   
#> presence of sulfides                                                                                
#>        |    |   |                                                                                   
#>        minimum horizon development                                                                  
#>             |   |                                                                                   
#>             daily under shallow water                                                               
#>                 |                                                                                   
#>                 other soils

cat(explainST("sapric frasiwassists"))
#> sapric frasiwassists
#> |      |    |   |                                                                                   
#> >25cm of sapric organic material                                                                    
#>        |    |   |                                                                                   
#>        not salty                                                                                    
#>             |   |                                                                                   
#>             daily under shallow water                                                               
#>                 |                                                                                   
#>                 soils with more than 30% organic matter content to a depth of 40cm or more

[dylanbeaudette]

Fat finger typo, thanks. I think that I had been using an older version of SoilTaxonomy and/or didn't see the suborder formative element. No action needed on this front.

I'll try to return to the meat of this issue this November.

[brownag]

In general it is a bit of a folly to expect formative elements to have consistent interpretation across taxonomy at least as far as specific diagnostics and depths and quantitative criteria are concerned. They are connotative, but not actually the definition.

Harmonizing formative elements and criteria was something that was interest to the Fundamental Changes task force, and that topic continues to be a recurring theme in questions I get in terms of "why" criteria are different between similarly-named taxa. The explanation is always the same: it is a "feature not a bug" (to allow e.g. different great groups with the same subgroup formative elements to differ in their usage). The scientific issue is of course that rarely do we have adequate data to statistically justify them being different. So while there are valid arguments that perhaps some taxa should be more similarly defined, they more or less are what they are.

I've always interpreted the goal of explainSMR as providing a simple (read: not adequate for actual diagnostics) explanation. If we want the actual explanation we read the criteria.

To be precise we would need to have a system that could allow for customization of any formative element for any taxon rather than just based on the taxonomic level it occurs at. This, I suppose, would be a layer on top of the existing system, not replacing it. We would not need to repeat redundant definitions, per the revisions made in 2021, we would just add additional clarification for taxa that needed it. Still it would be a significant job to do and I am not sure it is worth the effort. There is often not enough space in these explanation diagrams to give the level of detail needed to really clarify the differences. I think that after backfilling some of the missing ones and generalizing a bit several years ago this is functionality is adequate, with exception of the issues raised above in 2023.

"typic" -> "central theme of subgroup concept" and "hapl" as described in #55 (comment)

1. typic should say great group not subgroup
2. typic is not necessarily the "central theme"; this has shifted a bit in time probably, and the meaning of the word diverges from the functional definition, but with the way the modern keys are structured "typic" simply means "not anything that keyed out before it"; it need not be "central" in any particular way. KST13 p. 70, endnote 2, gives a great explanation of this but with far too many words for us to use.

There are probably a handful of explanations that could be shortened/simplified further, like references to soil moisture regimes--instead of saying the SMR class name we could give a less jargony explanation e.g. "xeric SMR" is more simply a soil in an area with a "dry summer"