5

I need to calculate green spaces in the metropolitan area for 155 cities around the world in QGIS.

Greenery would include everything from parks, trees, playgrounds, golf, vineyards, etc. Any advice where to start?

I have found a shapefile from NYU for metropolitan areas and now looking for each city under their departments, but already see an issue with that.

I wanted to do data scraping from OSM maps (it would be not all including). Also tried iTree even for approximate results but there were cities with missing or unclear satellite images.

Improve this question
1

2 Answers 2

Reset to default
5

Considering the wide scope of countries that you are looking at, OSM may well be your best bet.

If you are writing your own scripts, then you could load the 'world' file, filter for the cities, then filter for closed ways (ie. polygons) with tags that match your definition of 'greenery'.

An alternative approach would be to use 4 band (RGB,IR) satellite imagery if you can get it (big IF!). Then classify pixel areas according to 'NDVI' (Normalized Difference Vegetation Index) or 'EVI' (Enhanced Vegetation Index). This would also include things like gardens, grass verges (if your imagery is fine enough) etc.

Improve this answer
5
  • As for the imagery, the best resolution was 300m.
    – tji2019
    Commented Oct 29, 2019 at 13:52
  • Thank you! Is there already a script that could help me for OSM?
    – tji2019
    Commented Oct 29, 2019 at 13:53
  • 2
    I don't know of one, but you could look at Overpass for queries. You're going to want to have very specific queries, eg. one metro area at a time. And combine that with the greenery tags. The latter are documented on their wiki. Eg. leisure parks: wiki.openstreetmap.org/wiki/Tag:leisure=park?uselang=en
    – winwaed
    Commented Oct 29, 2019 at 13:55
  • Actually there is (at least for some regions) OSM-landuse-data.
    – Erik
    Commented Oct 29, 2019 at 14:05
  • Yes I think tji is going to have to query multiple tags and combine them. Different tags and definitions will work in difference places.
    – winwaed
    Commented Oct 29, 2019 at 14:34
1

As already mentioned using satellite imagery would be an option. Sentinel 2 Data might be useful since it has a 10m resolution in the visible spectrum and 10 or 20m in the NIR. There is an aws s3 requester pays bucket to access the data. I think if you register an aws account you have free data flat for your first year (see aws free tier). You can use the sentinel EO-Hub Browser to check which tiles you need.

There is also ready to use landcover data for European cities within the Urban Atlas project for download as shapefiles. I believe they use data from the SPOT satellites. Maybe you find a publication that describes the process how they did it and can apply it to other data sources.

Improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.