Mapping deforestation

classic Classic list List threaded Threaded
36 messages Options
12
Reply | Threaded
Open this post in threaded view
|

Mapping deforestation

Lorenzo Stucchi
Hi all,

I’m Lorenzo, the vice-president of PoliMappers, the YouthMappers chapter in Politecnico di Milano, we are going to organize a mapathon in Milan talking about deforestation, in collaboration with the Semillero Geolab UdeA, a Colombian chapter of YouthMappers.

We will would like to map land-cover in an area near the Amazonian forest, the mappers will be people that have few experience in mapping so we are thinking to map basic elements in landuse and we found this proposal [1] on map basic elements. 

The idea was to map: bareland, artificial surface and forest.

Should be a good idea to map following this idea?

Thanks for all the possible answer.

Best regards,
Lorenzo Stucchi 




_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Peter Elderson
you can use landcover, it has about 160K uses now by 6000 users, but you should know that
a. landcover is not currently rendered by OSM Carto. 
b. THe proposal states that "All areas in the landcover features of natural=* will be transfered to landcover=*. ...The tags landuse=grass, landuse=forest and landuse=railway will be transfered to landcover=*." This is certainly not going to happen any time soon.

If you decide to use landcover for existing area's, use it in addition to current mapping. For instance, if you replace landuse=grass with landcover=grass, or natural=wood with landcover=trees, most renderings suddenly will show these areas as grey. I would not recommend that, in fact you'll get a storm of protest! So in these cases, you use landcover in addition to landuse, and when landcover gets rendered (which I'm sure it will, some time) you can alter landuse without greying the map. 

To add patches of trees and grass in an area with landuse=residential, industrial or leisure, you can use landcover, though it will not yet show op on OSM Carto.

Also, major editors allow entering landcover, but have no special support for landcover tags.

Of course, if you make your own map style for the project, you can make landcover show up, just remember that the standard OSM Carto map does not.


Fr gr Peter Elderson


Op ma 11 mrt. 2019 om 15:16 schreef Lorenzo Stucchi <[hidden email]>:
Hi all,

I’m Lorenzo, the vice-president of PoliMappers, the YouthMappers chapter in Politecnico di Milano, we are going to organize a mapathon in Milan talking about deforestation, in collaboration with the Semillero Geolab UdeA, a Colombian chapter of YouthMappers.

We will would like to map land-cover in an area near the Amazonian forest, the mappers will be people that have few experience in mapping so we are thinking to map basic elements in landuse and we found this proposal [1] on map basic elements. 

The idea was to map: bareland, artificial surface and forest.

Should be a good idea to map following this idea?

Thanks for all the possible answer.

Best regards,
Lorenzo Stucchi 



_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Mateusz Konieczny-3
Mar 11, 2019, 4:32 PM by [hidden email]:
you can use landcover, it has about 160K uses now by 6000 users
6000 users? How you know that?


_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Paul Allen
On Mon, 11 Mar 2019 at 17:18, Mateusz Konieczny <[hidden email]> wrote:
Mar 11, 2019, 4:32 PM by [hidden email]:
you can use landcover, it has about 160K uses now by 6000 users
6000 users? How you know that?

I'm pretty sure overpass-turbo can do that.  I can't figure out how you ask it to do that, but at
least I've given you part of the answer.

--
Paul


_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Mateusz Konieczny-3



Mar 11, 2019, 6:29 PM by [hidden email]:
On Mon, 11 Mar 2019 at 17:18, Mateusz Konieczny <[hidden email]> wrote:
Mar 11, 2019, 4:32 PM by [hidden email]:
you can use landcover, it has about 160K uses now by 6000 users
6000 users? How you know that?

I'm pretty sure overpass-turbo can do that.  I can't figure out how you ask it to do that, but at
least I've given you part of the answer.
If it is done with Overpass turbo then I would love to see query, I tried something like that
in past and it worked only for very rarely used tags due to
performance issues.


_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Sergio Manzi
In reply to this post by Mateusz Konieczny-3
On 2019-03-11 18:17, Mateusz Konieczny wrote:
Mar 11, 2019, 4:32 PM by [hidden email]:
you can use landcover, it has about 160K uses now by 6000 users
6000 users? How you know that?

https://taginfo.openstreetmap.org/keys/landcover says:

"Objects with this key were last edited by 2 025 different users."

Cheers,

Sergio


_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

smime.p7s (4K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Mateusz Konieczny-3
In reply to this post by Lorenzo Stucchi
Mar 11, 2019, 3:15 PM by [hidden email]:
The idea was to map: bareland, artificial surface and forest.

Should be a good idea to map following this idea?
Mapping forests (or more accurately tree-covered areas) is useful and welcomed.

But two other categories are more complicated, in OSM areas are mapped
more accurately.

For example we are not mapping "bareland"/"artificial surface" but more specific
objects like

landuse=construction  (construction site)
landuse=farmland
landuse=meadow
landuse=railway
natural=beach
landuse=industrial
etc etc

is proposal from 2014 and is rarely used (not used) in mapping.

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Mateusz Konieczny-3
In reply to this post by Sergio Manzi



Mar 11, 2019, 6:44 PM by [hidden email]:
On 2019-03-11 18:17, Mateusz Konieczny wrote:

Mar 11, 2019, 4:32 PM by [hidden email]:
you can use landcover, it has about 160K uses now by 6000 users
6000 users? How you know that?

https://taginfo.openstreetmap.org/keys/landcover says:

"Objects with this key were last edited by 2 025 different users."

Cheers,

Sergio

What answers completely different question (it is possible that
tag was added by 200 users and then other people edited objects
with it).

Though given that even this value is so low makes 6000
people adding this tag even less likely.

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Tod Fitch
In reply to this post by Mateusz Konieczny-3

On Mar 11, 2019, at 10:44 AM, Mateusz Konieczny <[hidden email]> wrote:

is proposal from 2014 and is rarely used (not used) in mapping.


Maybe not that one, but I certainly map per https://wiki.openstreetmap.org/wiki/Landcover specifically I use it to map areas with trees, scrub and grass.

Cheers!



_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

signature.asc (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Christoph Hormann-2
In reply to this post by Lorenzo Stucchi
On Monday 11 March 2019, Lorenzo Stucchi wrote:
> [...]
>
> The idea was to map: bareland, artificial surface and forest.

As a general principle in OSM we map positively what verifiably exists,
not the lack of something.  What you call "bareland" is at odds with
that because it is essentially negatively defined (as land lacking
vegetation).  So in terms of choosing tags i would recommend you focus
on positively characterizing what you observe.

In areas of deforestation large areas of bare ground is typically a
transient state that will overgrow again rather quickly (within months
or a few years) with some form of vegetation, ob become regularly
maintained by humans in some form (like farmland).

We also do not have any established tagging for what you
call "artificial surface" because artificial surfaces are created
usually for specific purposes and we typically map them depending on
this purpose.  

Whatever tagging concept you choose you should document your plans and
discuss them with the local community beforehand to make sure your
plans are compatible with the mapping and tagging habits of the local
community.

--
Christoph Hormann
http://www.imagico.de/

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

marc marc
Le 11.03.19 à 19:27, Christoph Hormann a écrit :
> we map positively what verifiably exists

with different imagery, it is possible to verify:
- that an old one shows trees: old:landcover=trees
- that a more recent one shows no trees. landcover=xyz landuse=farmland
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Mateusz Konieczny-3
Mar 11, 2019, 8:58 PM by [hidden email]:
Le 11.03.19 à 19:27, Christoph Hormann a écrit :
we map positively what verifiably exists

with different imagery, it is possible to verify:
- that an old one shows trees: old:landcover=trees
- that a more recent one shows no trees. landcover=xyz landuse=farmland
We do not map what was in the past and disappeared.

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Peter Elderson
In reply to this post by Mateusz Konieczny-3
Sorry, 2000. 

Vr gr Peter Elderson


Op ma 11 mrt. 2019 om 18:18 schreef Mateusz Konieczny <[hidden email]>:
Mar 11, 2019, 4:32 PM by [hidden email]:
you can use landcover, it has about 160K uses now by 6000 users
6000 users? How you know that?

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

marc marc
In reply to this post by Mateusz Konieczny-3
Le 11.03.19 à 21:36, Mateusz Konieczny a écrit :

> Mar 11, 2019, 8:58 PM by [hidden email]:
>
>     Le 11.03.19 à 19:27, Christoph Hormann a écrit :
>
>         we map positively what verifiably exists
>
>
>     with different imagery, it is possible to verify:
>     - that an old one shows trees: old:landcover=trees
>     - that a more recent one shows no trees. landcover=xyz landuse=farmland
>
> We do not map what was in the past and disappeared.

who is we ? check taginfo for ex 288k old_name or some lifecycle prefix
if some users keep old name, I see no issue to keep the old cover in
addition to the new one.
_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Sergio Manzi
In reply to this post by Mateusz Konieczny-3
On 2019-03-11 18:47, Mateusz Konieczny wrote:
Mar 11, 2019, 6:44 PM by [hidden email]:
On 2019-03-11 18:17, Mateusz Konieczny wrote:
Mar 11, 2019, 4:32 PM by [hidden email]:
you can use landcover, it has about 160K uses now by 6000 users
6000 users? How you know that?

https://taginfo.openstreetmap.org/keys/landcover says:

"Objects with this key were last edited by 2 025 different users."

Cheers,

Sergio

What answers completely different question (it is possible that
tag was added by 200 users and then other people edited objects
with it).

... and that doesn't change things very much: it means that even if only 200 (or 20) users have originally added those tags, 2025 users who have last edited those objects have somehow agreed with the tag being there, which I think is the important thing and probably that's why Taginfo exposes that statistic.


Though given that even this value is so low makes 6000
people adding this tag even less likely.


As you have probably seen Peter has amended the number (which initially was little a bit "optimistic") to something very comparable.

Cheers,

Sergio


_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

smime.p7s (4K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Mateusz Konieczny-3
In reply to this post by marc marc



Mar 11, 2019, 10:28 PM by [hidden email]:
Le 11.03.19 à 21:36, Mateusz Konieczny a écrit :
Mar 11, 2019, 8:58 PM by [hidden email]:

Le 11.03.19 à 19:27, Christoph Hormann a écrit :

we map positively what verifiably exists


with different imagery, it is possible to verify:
- that an old one shows trees: old:landcover=trees
- that a more recent one shows no trees. landcover=xyz landuse=farmland

We do not map what was in the past and disappeared.

who is we ? check taginfo for ex 288k old_name or some lifecycle prefix
if some users keep old name, I see no issue to keep the old cover in
addition to the new one.
old_name is OK - it is for name that is older/outdated but still in use among
some people

The only case where it is a bit OK is for recently changed features
(note="cycleway tagged here is not yet visible on aerial images"
or note-"house visible on aerial images is demolished as of 2019-03").

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Mateusz Konieczny-3
In reply to this post by Sergio Manzi



Mar 11, 2019, 11:38 PM by [hidden email]:
On 2019-03-11 18:47, Mateusz Konieczny wrote:
Mar 11, 2019, 6:44 PM by [hidden email]:
On 2019-03-11 18:17, Mateusz Konieczny wrote:
Mar 11, 2019, 4:32 PM by [hidden email]:
you can use landcover, it has about 160K uses now by 6000 users
6000 users? How you know that?

https://taginfo.openstreetmap.org/keys/landcover says:

"Objects with this key were last edited by 2 025 different users."

Cheers,

Sergio

What answers completely different question (it is possible that
tag was added by 200 users and then other people edited objects
with it).

... and that doesn't change things very much: it means that even if only 200 (or 20) users have originally added those tags, 2025 users who have last edited those objects have somehow agreed with the tag being there, which I think is the important thing and probably that's why Taginfo exposes that statistic.

"2000 users" and "2000 people not disliking it enough to remove all noticed instances"
is a bit different.

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Sergio Manzi
On 2019-03-12 00:00, Mateusz Konieczny wrote:
Mar 11, 2019, 11:38 PM by [hidden email]:
On 2019-03-11 18:47, Mateusz Konieczny wrote:
Mar 11, 2019, 6:44 PM by [hidden email]:
On 2019-03-11 18:17, Mateusz Konieczny wrote:
Mar 11, 2019, 4:32 PM by [hidden email]:
you can use landcover, it has about 160K uses now by 6000 users
6000 users? How you know that?

https://taginfo.openstreetmap.org/keys/landcover says:

"Objects with this key were last edited by 2 025 different users."

Cheers,

Sergio

What answers completely different question (it is possible that
tag was added by 200 users and then other people edited objects
with it).

... and that doesn't change things very much: it means that even if only 200 (or 20) users have originally added those tags, 2025 users who have last edited those objects have somehow agreed with the tag being there, which I think is the important thing and probably that's why Taginfo exposes that statistic.

"2000 users" and "2000 people not disliking it enough to remove all noticed instances"
is a bit different.

Right


_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging

smime.p7s (4K) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Joseph Eisenberg
In reply to this post by Mateusz Konieczny-3
> “We will would like to map land cover in an area near the Amazonian forest”

I’d recommend that you start by mapping the existing forested areas with natural=wood or landuse=forest, and areas of water with natural=water and water=lake / =river, or natural=wetland for swamps, marshes, mangroves, bogs, etc.

If you can clearly identify other types of vegetation (from aerial imagery, I assume?) there are several specific tags that can be used. You should not tag something non-specific like “clearing” or “bare ground”.

Natural=scrub is used for areas mainly covered shrubs or bushes, eg woody plants about 1m to 3m in height, natura=heath is for dwarf shrubs (probably only found in the high Andes on Colombia), natural=grassland can be used for alpine areas above the treeline (though these may also be heath or wetland=bog). Areas of rocky land without vegetation can be natural=bare_rock, =sand, =scree, =shingle

The most relevant tags for landcover with heavy human use in Colombia might be landuse=orchard for palm oil, banana, and coffee plantations, landuse=farmland for seasonal or annual crops, landuse=meadow for pasture and hay fields, landuse=residential / =industrial for developed areas with houses / industry. 

The idea should be to map as specifically as possible. If you are not certain of the type of vegetation or landuse for a certain area, then leave that place untagged for now. Someday a local person can add the proper tagging by visiting that place in person.

Re: the landcover tags. These have not been approved by the community, though there is a small, vocal minority that wants to use them instead of the established tags that start with landuse= and natural=. I would not use these, because there are much more widely used equivalents that are approved or in use for many years.

Don’t use landcover=trees; use natural=wood or landuse=forest. 

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
Reply | Threaded
Open this post in threaded view
|

Re: Mapping deforestation

Christoph Hormann-2
In reply to this post by Peter Elderson
On Monday 11 March 2019, Peter Elderson wrote:
> Sorry, 2000.

IIRC the saying is "two wrongs does not make a right".

Original use of tags with the landcover key, that is mappers creating a
new geometry with a landcover tag, is as follows (based on data from
2019-02-28):

72848 ways/relations (more of half of these created in organized mapping
with tagging not being the free choice of the mapper)
1310 different users
494 of which have used the key exactly once (this, i.e. that about 1/3
to half of the genuine active users of a tag have only used it once is
pretty standard but still this has to be kept in mind when
contemplating such numbers)

The reason why taginfo reports only the number of users who have last
touched features with this key is not because this is particularly
meaningful information but because this can be counted quite easily
when processing a planet file (which is what taginfo does on a daily
basis) while numbers on active users (i.e. who maps features with a tag
or who adds a tag to features) can only be determined from the history.

I can highly recommend Frederik's talk on the matter of OSM statistics
which discusses this in detail:

https://www.youtube.com/watch?v=Kx0KuvkbvfQ

--
Christoph Hormann
http://www.imagico.de/

_______________________________________________
Tagging mailing list
[hidden email]
https://lists.openstreetmap.org/listinfo/tagging
12