[Tagging] From the Australian fires (still burning unfortunately) how to map burnt areas
joseph.eisenberg at gmail.com
Mon Jan 27 01:47:50 UTC 2020
For woods the problem is that there are 3 different levels of fire
impact, and only the most severe is clearly visible in aerial imagery
or from a distance - but this is also quite dangerous to survey in
) 1A "crown fire" (US English) burns the tallest trees which comprise
the canopy or "crown" of the woodland. In the Northwest USA, this
would be the tall Ponderosa pines, Douglas firs or similar very tall
trees. This sort of fire leaves standing stags with no needles or
leaves and the ground is usually scorched black - very obvious.
2) But most forest fires rarely reach the crown: most of the fire will
be limited to the understorey (shrubs and immature trees). This is not
always obvious on aerial imagery
3) Many areas of the fire will only burn the ground-level needles,
leaves and herbs. This is not going to be apparent from the air.
Areas of scrub can also have the last type of fire, though it is more
common that the "crown" or "canopy" of the shrubland is burned, since
it is closer to the ground and shrublands tend to have a very dry
Dwarf shrub and heath and grassland areas will usually only have the
last type and it will be quite obvious, but the herbs will regrown
during the first rainy / growing season after the fire, so the burn is
very transient: from just a month or two in wet tropical areas to
perhaps a max of 9 months in arid areas.
Similarly, scrubland areas (and burnt undergrowth) will have new
growth of grasses and herbs within months and regrowth of woody shrubs
within a couple of years at the most.
So if it reasonable to map this? If you do map it, you had best be
prepared to remap the area in 12 months and again after a couple of
years, or the data will be quickly incorrect.
I would also think it is pretty necessary to have the current natural
vegetation mapped at the same time for surrounding areas, or the
extent of the burned area will lack context.
- Joseph Eisenberg
On 1/27/20, Alan Mackie <aamackie at gmail.com> wrote:
> There are some tags documented on the Russian wiki page for key:wood 
> that describe various types of damage to wooded areas, one of which is
> wood:damage=burnt. I don't think it's been used much outside of Russia, but
> it seems fairly reasonable for areas likely to regenerate.
> There have been other disaster tagging schemes that seem to be used on
> occasion, but they always seem to be listed under a particular disaster's
> wiki page rather than in some centralised location.
> It would be nice to get some more consistency here, but I think the problem
> with this sort of tagging is that these tend to be the sort of thing that
> should be locally surveyed and updating OSM generally isn't the priority
> unless it is somehow part of a specific organisation's workflow.
> : https://wiki.openstreetmap.org/wiki/RU:Key:wood
> On Sun, 26 Jan 2020 at 22:27, Warin <61sundowner at gmail.com> wrote:
>> I have come across a German mapper who has used 'landuse=brownfield' to
>> map some recently burnt areas in Australia.
>> I know this is not appropriate as it is not a land use, nor does it meet
>> the OSM meaning of 'brownfield' in all situations.
>> Note: this is done in areas, no matter if it is a farm field, recreation
>> ground, residential areas etc.
>> Tagging buildings and other OSM features is easy where they are wholly
>> burnt using the life cycle tagging.
>> Should this data be entered in OSM?
>> Are flooded damaged areas mapped? Earthquake damaged areas?
>> If entered how should they be tagged?
>> The keys 'landuse', 'landcover', 'natural' don't look usable to me.
>> It may not be possible to use a sub key as the burnt areas do not
>> conform to the presently mapped features.
>> If mapped then the tagging should be inclusive of other disasters
>> (natural or not).
>> Tagging mailing list
>> Tagging at openstreetmap.org
More information about the Tagging