All looks excellent to me.
Is it possible to put a note on the boundary asking people not to edit it's shape? Or is this not encouraged by OSM?

Thank you so much! I am in debt and if there's anything you need me to do in future, please give me a shout.
Yes I'm happy to make these amendments either tonight or tomorrow. I want to make sure I familiarize myself better to avoid making further mistakes.
Thanks again to both of you.

I don't want to go into the politics of it because it's mid-blowing. But theoretically, if the official documents of a boundary says 'it follows the centreline of a river', then legally, the boundary can be changed when the river changes. That's what the Lake District has done. But the Brecon Beacons has followed the original river centre line to keep it straightforward. The Brecon Beacons boundary is only as good as MasterMap, which indeed is undergoing a major improvement excercise where much of it's upland data digitised at 1,10k scale, will be redigitsed to 1:2500 scale. In this case, the Brecon Beacons has taken the apparoach that it will continue to improve the boundary so that it's in alignment with MasterMap (this excludes old rivers, roads etc).

Although the OS are not boundary designators, they have worked with the boundary commission etc, and they are indeed the official source for political boundaries - and this is reflected in their MasterMap product only. National Park boundaries and maybe some others are boundaries that OS no not represent in their Mastermap product (yet).

Yes you're absolutely right. The boundary was first created in 1955 and in some lowland areas the river has moved quite a bit since then. But for upland streams (like the one I shared a link) these are in steep valleys and are unlikely to move much. I'm not saying we will be able to copy the National Park boundary in all areas an assume its correct, as it follows old roads in areas which have since been straightened, but much of it can be traced.

Agreed. So what's the next step for committing the new national park boundary and removing the old?

We've also made some of our datasets open data, such as Public rights of ways, Fforest Fawr Geopark, and International Dark Sky core.

https://data.gov.uk/publisher/brecon-beacons-national-park-authority

Natural Resources Wales has also made their LiDAR data free and from this there's potential to calculate building heights as an additional attribute for every building in Wales, and with EA LiDAR, England too! The potentials are huge :-)

This is going to sound really geeky of me but I'm looking forward to improving steams and rivers like this one, although I OS has released an open roads and open rivers dataset , and if more accurate, I expect this may be used to supersede all roads and rivers anyway:

changeset/39117825#map=17/51.75817/-3.40991

You can see them all along there Herefordshire boundary from Hay-on-way if you pan in a southerly direction. This is a bizarre section as it zig-zags - but that's what the offical boundary does. It has just been generalised in open data boundaryline:

node/4170092739#map=18/52.03392/-3.09150

They are tiny inaccuracies really so nothing of major concern for illustrative purposes - but may need amending if people use it as 'data'.

I notice the English/Welsh border is inaccurate in some areas too. Another benefit of keeping the new national park boundary.

Thank you The Maarssen Mapper, that's all that I want. The boundary-line data has already been generalized by OS. In mastermap, there is a boundary data which are the 'official political boundaries'm and these are what government use. Of course we're not allowed to follow that data directly as it breeches copyright, but as the National Park is Opendata, there's no reason why we can't trace this.

I think the best way forward is to remove the 2 changesets I created in error, and replace the new boundary with the old. I'm happy to to do this with pointers, but also I would appreciate help.

The National Park boundary I have uploaded is the official boundary. In some areas it follows roads, rivers, and political boundaries from OS MasterMap. Therefore what has already been digitised in OSM is already 'garbage.'

I have a detailed knowledge about the national park boundary and what it follows, and I would love to learn the system more so that I can start aligning rivers and roads and political boundaries so that they too can become really accurate. Is this not a good way forward?

No unfortunately I wasn't aware that it existed. In fact I wasn't aware that I was uploading to the live copy, but thought I was uploading to a trace. It has been a steep learning curve. Nevertheless unfortunately the system is full of flaws and limitations and I'm at the process of learning these limitations. I agree, the boundary needs some generaliaing- but surely the goal here is to create more detailed and more accurate data? That's all I've tried to do, and generalising data to the stage where it will become inaccurate seems bizarre,

I apologise for causing inconvenience, but we all got to start somewhere, and this is a learning curve for me. Encouragement is what's needed, to gain more data editors.

Ah yes I see. Here they are a direct copy of the 'official' administrative boundaries from Ordnance Survey. I won't carry out any new changeset uploads. But if it would help the situation, I can generalise the boundary if needed?

If only OSM had decided to partner with QGIS or something rather than creating their own dodgy and complex interfaces. Maybe this would be far more straightforward!

As a separate note, all Welsh National Park boundaries have been granted Open Data, and I believe England and Scotland has followed. So learning from what I have done here and coming up with a suitable method, there may be an opportunity for me to work with the other 14 National Parks so that they can upload the correct versions of their boundaries to OSM.

Thank you both, lesson learnt and apologies for the inconvenience. My initial aim was to upload it as a trace and amend the existing boundary - but I had issues with the web interface and then I thought this was what I was doing with JOSM.

Each Way in relation/6194617 is tagged as' Brecon Beacons National Park' because I didn't follow existing Ways and I weren't sure of what to tag my 'new' Ways as. Each of these Ways is 2,000 nodes or less each so hopefully it will work.

The boundary does follow waterways and roads in some places, so Maaseen you are correct in maybe I should have taken the slower approach and use existing Ways in some places (I would have done this if I successfully managed to upload the boundary as a trace).

But on the other hand, where the National park boundary follows roads and rivers, it is more accurate than the existing Ways. So maybe another approach would be to amend some roads and rivers to follow the boundary Way?

I am an amateur at OSM so I'm all ears to what you think would be the best approach.

Thanks again!

Thank you! I now need to remove this change set and my other accidental changeset: changeset/39120226

Do you know how these can be removed?

Thanks again!

Shaun

And yes, it should be replacing relation/357283.

I can upload and a polyline instead? But I'm unsure how OSM handles the national park 'fill colour' if it's a polyline.

Hello. I'm an amateur at this and could do with some help. This was the only way I could upload the file (GPX traces upload kept failing). My goal is to replace the existing national park boundary (which has been digitised inaccurately) with this boundary. The problem with this boundary is that I've had to split it into multiple polygons to get around the 2,000 node limit.

This is what the boundary should be:
https://data.gov.uk/dataset/brecon-beacons-national-park-boundary

Thanks

The official and accurate boundary for the Brecon Beacons National Park. The polygon has been split into many parts limited to 2,000 nodes/vertices each.