[Talk-GB] Mapping of Kielder Forest(s)

[Russ Garrett]

I realised that while the open data doesn't seem to be good enough for
mapping the detail of the forest, it probably is still useful for the
outline.

So I've added it as a multipolygon here (tagged with landuse=forestry):

https://www.openstreetmap.org/relation/13163840

I've applied some simplification to this (there are a few places where
only roads are part of the forest outline), and I've attached the
boundary to the outline of Kielder Water/Bakethin Reservoir to avoid
duplication round there.

I think it should be useful, if only for identifying where there might
be fences. I think in some locations the geometry of this outline is
obviously a bit offset and it can be manually corrected, especially
where it coincides with residential landuse.

Hopefully this one is useful at least, and at some point we can work
out how to render it sensibly...

Russ

On Fri, 3 Sept 2021 at 20:54, Russ Garrett <russ at garrett.co.uk> wrote:
>
> While looking into this earlier this week, I ended up down a bit of a
> rabbit hole. The Forestry Commission has a rather interesting
> collection of open data (ODbL-licensed, so compatible with OSM in
> theory).
>
> Most relevant to this discussion, they have a map of
> "subcompartments", which I think basically equates to landcover -
> they're all labeled with type of land use, species, planting date,
> etc.
>
> I started looking at this for Kielder, and first impressions are quite
> good (these PNGs are all quite large, you might struggle opening them
> on mobile):
> https://f001.backblazeb2.com/file/russss/kielder-subcompartments-by-date.png
>
> This dataset is last updated in 2019, which is at least roughly
> contemporaneous with most of the mapping in OSM, so here are the OSM
> woods/forests overlaid on that map:
> https://f001.backblazeb2.com/file/russss/kielder-subcompartments-osm-overlay.png
>
> The Forestry Commission's mapping of gaps/cutlines is quite
> inconsistent. Some are shown, some aren't. It doesn't seem to be
> related to compartment boundaries. It doesn't really match the OSM
> boundaries that well in many places either.
>
> But most notably, there are some areas where it's clearly wrong - like here:
> https://f001.backblazeb2.com/file/russss/Screenshot-2021-09-03-at-20.15.08.png
>
> That's OSM on top of the Forestry Commission data on top of Bing
> there. I'm assuming the Bing imagery dates from after 2019 (copyright
> date is 2021 in most places).
>
> The forested area at the top right, which clearly looks like it's
> coniferous and planted well before 2019, is mostly marked as "acid
> grassland", as are most of the (apparently recently-planted or at
> least recently-felled) areas in the middle.
>
> The forested area to the left of the image is marked as "felled" but
> the imagery is showing a forest which looks a lot older than 2 years
> to me.
>
> Not much of a conclusion here other than "maybe this data isn't as
> useful as it looks". It's pretty though.
>
> Cheers,
>
> --
> Russ Garrett
> russ at garrett.co.uk



-- 
Russ Garrett
russ at garrett.co.uk