Thu 5 Dec 2019 10:43PM

Adding trees

SL Stuart Lowe Public Seen by 66

The Leeds Ecosystem, Atmosphere & Forest (LEAF) Centre at the University of Leeds recently did a campus tree survey and mapped 1400 individual trees (not on Open Street Map). They then used the species/genus/height/diameter data to work out the environmental benefits such as CO2 storage, pollution reduction, and flood mitigation using the iTree Eco tool e.g. https://leaf.leeds.ac.uk/news/i-tree-leeds-putting-a-value-on-the-citys-trees/

Their existing data was collected from paper-based surveys that they did from their own maps so there is potential for them to add their existing dataset to OSM. I've suggested they do that and it would be good to know if anyone has experience adding a bulk dataset as I don't.

I've also suggested that they broaden their reach to a wider community and get people to add trees directly to Open Street Map so that the data are available to everyone and not locked up in a silo (there are many tree data silos). I realised that it might be good to make it easier for people to add trees (and the relevant information needed for environmental modelling) to OSM so I've started making a web app that will be just for this task. Obviously people can also use the existing web editors etc but I wanted to provide a helping hand for various aspects of this.

I have made a start using the dev version of OSM. I found a Javascript OSM/OAuth library that means I can get the app to log in (but logging in again later seems to make a new app authorisation). I've worked out the basics of using Overpass to get nearby trees. I'm part way to constructing the appropriate changesets in XML. I'm also making some tools to help ordinary people with adding tree height and identifying the tree species/genus. I've done a few things already:

Do other people think this is useful? Is this a bad way for me to go about this?

My main aim is to get more trees mapped and their data available. But I also want individual users to get credit for the data they add (hence needing to work out how to do OAuth in the browser).


Stuart Lowe Tue 10 Dec 2019 10:19AM

In addition to this, The Open University are trying to create a standard for recording individual trees in surveys. Is there any way to feed this into adding additional tags to natural=tree so that an OSM survey would be compatible?


Brian Prangle Wed 11 Dec 2019 9:27PM

Hi Stuart I imported Birmingham City Council's tree database (several thousand data points) and learnt a lot from my mistakes -happy to advise


Brian Prangle Wed 11 Dec 2019 9:28PM

Stuart It would be great if the OU could discuss this directly with the chapter - do you have a contact?


Stuart Lowe Wed 11 Dec 2019 10:15PM

Brian, unfortunately I don't have any contacts with them. The OU TreeZilla account tweeted the survey and that was how I found out about their proposed standard https://twitter.com/Treezilla_org/status/1200064442425397248


Brian Prangle Wed 11 Dec 2019 11:25PM


The twitter feed sends me to a map based on GoogleMaps called treezilla.org ( http://treezilla.org ) but there is also a mention of a project called opentreemap which is a global opensource project whose data is released under the terms of the GNU Affero General Public License version 3. Not sure if this is compatible with ODbL. Are you working separately from or together with this initiative? If together, which one treezilla or pentreemap? It looks like there's some synergy here that might be worth exploring




Stuart Lowe Tue 17 Dec 2019 3:10PM

The very long OU survey is at http://www3.open.ac.uk/forms/individualtreedatastandard/

I'm not involved with either group. I was unable to find a link to open data on the OpenTreeMap site or the TreeZilla site. Do you have a link to a download (and the licence for the download)? OpenTreeMap and TreeZilla aren't the only data collection silos (e.g. The Woodland Trust have their own separate app for recording trees) and I was hoping to suggest that OSM be a good place to do this work in the open to avoid duplication.


Christian Ledermann Wed 18 Dec 2019 9:26AM

I have some notes on tree detection from aerial imagery


RobJN Tue 24 Dec 2019 4:22PM

The webapp idea is really great Stuart. Would certainly help people who want to focus on trees add this data to OpenStreetMap. At what stage do you start approaching the existing groups to bring them on board? Happy to help explain the benefits of OpenStreetMap to them if needed.


Huw Diprose Wed 29 Jul 2020 11:35AM

Hey Brian, really interested in the work you mentioned about importing a tree dataset from the council.

Do you know anything more about general availbility of those datasets? Are they things councils generally have as part of their environmental services, or are there legal responsibilities to collect that data?

or is it just the more progressive digitally enabled folks who are making them and putting them out into the world?

Would love to chat about how you found the data to use. And could likely help with a web app if that's the direction the community wanted to go in?


rskedgell Fri 7 Aug 2020 8:13AM

There's a dataset for Greater London's trees collated from local borough information by the Greater London Authority and released under OGL v3. It includes data from boroughs who never release any open data directly.

I'm gradually starting to match some trees in parts of the London Boroughs of Newham and Waltham Forest with Bing aerial imagery to get an idea of how accurate they are, before contemplating and import. The trees in at least one borough in south east London have already been imported by another user.



Brian Prangle Thu 6 Aug 2020 6:02PM

Hi Huw happy to chat offline contact me at community@mappa-mercia.org rgds Brian


joeldn Mon 29 Mar 2021 10:42AM

Hi all, currently i'm attempting to map a randomised gridded subselection of ancient woodland: 40*10m squares which make up 1% of the grid; despite being a tiny sample, it may reveal some clues about the wood's ecology. Using OSM for a randomised tree survey may be a thorny issue (quite literally at times), however the shared tools and methods are mostly perfect for conservation fieldwork. One tool is having problems however...

...when i stop walking to photograph and locate a tree, GPS accuracy falls dramatically on both direction and radius (screenshot attached for reference) - and this is in spring before the leaf canopy has appeared, which may further impair GPS.

using armchair mapping alongside surveys, i'm trying to cross-reference photo locations against lidar, but I would be keen to hear about any recommendations for improving GPS accuracy, would one of the osmuk camera hires be worth a try for example?

the randomised sampling method used is also available for comment.


RobJN Sun 11 Apr 2021 6:27PM

Hi @joeldn. What are you using as your GPS device? Mobile phones can be a bit hit and miss - a proper GPS device might be better but even then it is a challenging environment to get a good accurate location at slow speeds.

Having extra photos may help you to work out the location a bit better but maybe it just gets too confusing when in a large wood! Either way I see no harm in trying the TrekView Trekker Loan program linked to via the osmuk.org website.


joeldn Wed 14 Apr 2021 9:48AM

@RobJN thanks. The device is a mobile running vespucci; it seems fine on the move, but unsuitable for a site survey. foresters have various methods at their disposal: sometimes they are in a team of three: one holding a trimble, one putting down corner flags and one taking notes; sometimes if it's a single person, they try to locate the tree nearest the south-western corner, tie a ribbon to that one, then walk north to the tree nearest the north-west corner, tie another ribbon ... and so on. another interesting thing i learned is that they generally don't survey trees slimmer than a baked bean tin.


SK53 Wed 21 Apr 2021 4:22PM

@joeldn Read George Peterken's experience in Lady Park Wood about all the vagaries of detailed tree survey in ancient woodland. Personally I would use RTK to localise your quadrats & a traditional tool like a tape measure :-). With one good reference point, or better two, simple triangulation with tape measurements should be good enough. There are some dedicated (paid for) tree survey apps more designed for foresters (e.g., estimation of metres cubed of timber). Most EU countries do standard national forest surveys on a 10-15 year timescale on standard quadrats, spaced on a grid a few kilometres apart. I have the details of the Spanish one somewhere around.

I've never really got reliable enough measurements for accurate tree location using retail GPS or mobile phones, even in the open.


joeldn Fri 23 Apr 2021 7:09AM

This is really interesting @SK53, the methods for surveying trees really are quite sophisticated and...triangulation from a known point of reference sounds worth a try - the known point in this case might be a certain corner on a path that has lots of gpx traces - does that sound feasible?


Nick Ananin Thu 22 Apr 2021 10:19AM

For surveying trees under a canopy, GPS alone won't do it. Solutions then depend on your budget - either a) traditional tape and possibly surveyors tools ( e.g. Dumpy level) or b) forestry laser finders for which you still need line of site but you can use offsets. All comes down to precision and accuracy vs time and cost.
One thing people might consider is tagging individual trees (with the owners permission) for validation and monitoring.
Either way, there is a fantastic opportunity for education, both in terms of mapping but also mathematics (including trigonometry - including measuring heights), statistics, botany, soils etc. (you can see I was a forester!). The key is to make more of this sort of data available as 'open data'


joeldn Fri 23 Apr 2021 8:30AM

@Nick Ananin Re tagging trees, it is worth considering: there are the round aluminium discs with digits, then there are softer larger aluminium ones you can mark with a biro, which could be good for adding useful info, although they seem much less durable. Physical tags would be especially useful in the absence of metre accurate gps as it would make it easier for others to double check coordinates.