Seth, I'm not sure it's a fair test.
By way of piling onto Seth's point, though: When MMI did a vocabulary mapping workshop a long while ago, by far the least mappable set of vocabularies was habitat, which I think it is a fair analog to area_types. The problems (sic) with 'habitat' were that it was idiosyncratically categorized (what it looks like, what lives there, what physical parameters abound, what the 'primary purpose' was, ...), ordered (flat, hierarchical, network/mesh), and derived (lots of papers with new ordering concepts based on complex high-order principles; few primary classifiers; few meetings among all those classification system developers). Naming inconsistencies was a fourth-order problem!
So I endorse Seth's example, and IF the charge is to come up with a good all-purpose system, I think this is not it, nor does it move toward it. But if you want a solution that some/many people can use for some basic purposes, this is a good interim path.
(At the risk of sounding like a broken record: A good all-purpose system might be one in which the rule is not "please use terms from our vocabulary", but "please use uniquely identified terms from a vocabulary that follows these basic principles (has unique identifiers, is publicly available, blah blah blah)." That would provide a significant amount of interoperability, and no or little maintenance/review cost for the CF community.)
John
On Sep 26, 2011, at 15:56, Seth McGinnis wrote:
> Jonathan Gregory <j.m.gregory at reading.ac.uk> wrote:
>>> Do we get enough benefit by "standardizing" them
>>> [area-type names] to offset the cost in time and trouble
>>> of the growth of yet another complex name hierarchy? (I
>>> know. Some people will say "Yes!" I just have to ask.)
>>
>> It's a fair question. I am one of those who would say
>> "Yes"! If it turns out that this becomes a large problem
>> which we can't deal with effectively, we will have to think
>> again. So far that has not happened.
>
> Hi all,
>
> I would be one who says "No"; I think the area_types system is
> too restrictive, and that we should move away from standardized
> names for surface classifications. (I do have a replacement
> scheme in mind that I would like to float, once I'm not
> frantically helping revise a paper.)
>
> But rather than argue the case abstractly, let me propose a test
> case that I think will allow us to answer the question
> *experimentally*. I actually have data on hand that I've been
> pondering how to properly represent under CF.
>
> These are the different land-type classifications used by the
> land-surface models (NOAH, BATS, and MOSES) that are part of the
> regional climate models we work with. I don't know much about
> these classes other than which model they're coming from, and
> that they should be stored with a standard_name of "land_cover".
>
> Attached below are the 93 unique values for surface types from
> these 3 models. If it looks like we could hash out the details
> and extend the existing area_types list could be extended to
> cover this list in fairly short order, then I think we can say
> that the current system will be sufficient. If it's a bigger
> job than that, then I don't think it's adequate.
>
> Does that seem like a fair test? (And note, there's a fourth
> land-surface model that I don't have data from. I expect that
> when I do get it, I'll have another list of around 20-40 land
> types to reconcile with / add to the existing list. So proposed
> additions will likely come in big chunks like this, not one or
> two at a time.)
>
> Cheers,
>
> --Seth McGinnis
>
>
> Land-cover types used by land-surface models in NARCCAP:
>
> arable cropland
> bare ground tundra
> barren or sparsely vegetated
> bog or marsh
> cane sugar
> coffee
> cotton
> crop/mixed farming
> cropland/grassland mosaic
> cropland/woodland mosaic
> deciduous broadleaf forest
> deciduous broadleaf tree
> deciduous needleleaf forest
> deciduous needleleaf tree
> deciduous shrub
> deciduous tree crops (temperate)
> dense deciduous broadleaf forest
> dense deciduous needleleaf forest
> dense drought deciduous forest
> dense evergreen broadleaf forest
> dense mixed evergreen + deciduous forest
> dense needleleaf evergreen forest
> desert
> dry farm arable
> dryland cropland and pasture
> dwarf shrub (tundra transition + high altitude wasteland)
> equatorial rain forest
> equatorial tree crop
> evergreen broadleaf
> evergreen broadleaf cropland
> evergreen broadleaf shrub
> evergreen broadleaf tree
> evergreen broadleaf woodland
> evergreen needleleaf
> evergreen needleleaf tree
> evergreen shrub
> forest/field mosaic
> grassland
> herbaceous tundra
> herbaceous wetland
> ice
> ice cap/glacier
> inland water
> irrigated crop
> irrigated cropland
> irrigated cropland and pasture
> maize
> mangrove (tree swamp)
> mixed dryland/irrigated cropland and pasture
> mixed forest
> mixed shrubland/grassland
> mixed tundra
> mixed woodland
> nursery + market gardening
> ocean
> open deciduous broadleaf woodland
> open deciduous needleleaf woodland
> open drought deciduous woodland
> open mixed evergreen + deciduous woodland
> open needleleaf evergreen forest
> open tropical woodland
> open water
> paddy rice
> pasture + shrub
> pasture + tree
> rough grazing + shrub
> sand desert + barren land
> savanna
> semi arid rough grazing
> semi desert + scattered trees
> semi-desert
> short grass
> shrub desert + semi desert
> shrubland
> snow or ice
> tall grass
> tea
> temperate meadow + permanent pasture
> temperate rough grazing
> thorn shrub
> tropical broadleaf forest (slight seasonality)
> tropical grassland + shrub
> tropical pasture
> tropical savanna (grassland + tree)
> tundra
> urban
> urban and built-up land
> vineyard
> water and land mixture
> water bodies
> wooded tundra
> wooden wetland
> woodland + shrub
> _______________________________________________
> CF-metadata mailing list
> CF-metadata at cgd.ucar.edu
> http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
John Graybeal <mailto:jgraybeal at ucsd.edu>
phone: 858-534-2162
Product Manager
Ocean Observatories Initiative Cyberinfrastructure Project:
http://ci.oceanobservatories.org
Marine Metadata Interoperability Project:
http://marinemetadata.org
Received on Mon Sep 26 2011 - 19:12:20 BST
