How green are London’s streets? (part I)

In February this year, using the nifty i-Tree tools, the London Tree Officers Association announced that their studies showed that London has 21% tree coverage. Inspired by this, browsing through the Greater London Authority’s London Data Store for landscape related datasets, I found a 2011 estimate of street trees per borough.

The dataset contains street tree estimates for all the london boroughs, excluding TfL maintained streets (red routes) and presumably Highways Agency maintained motorways. Could I use this data to create a league table of how green London boroughs streets are?

The GLA have attempted to estimate this in their report Branching Out: The future for London’s street trees as part of the Mayor’s larger RE:LEAF strategy, however, they only estimate how many trees by square kilometre or by using the Department for Transport’s total borough road length data.

Street trees per kilometre (London Assembly, 2011)

The report highlights some issues with this – the road length data includes roads that were not part of the tree estimates. To get a better idea, I’d have to crunch some of the data myself, starting with a couple of boroughs as a test.

Firstly, I’ll have to make some assumptions about what a street tree is and how to interpret the dataset:

  • Street trees occur alongside roads either in a nib planting, pavement, roadside verge or other road feature like a central reservation or roundabout.
  • Street trees occur on urban / suburban roads only (otherwise estimates would have to include hedgerow trees and wooded areas).
  • The dataset excludes trees in parks, motorways, TfL managed and other private roads.

Using OS OpenData, especially VectorMap District and Meridian 2 map data in QGIS, I clipped the road data to the borough outlines and settlement areas (important to exclude rural lanes on the edge of Greater London), and excluded roads classified as motorways, tunnels, and private roads. Next is to add an extra column to the road data’s attribute table and get QGIS to populate it with the road lengths. Then I classified the TfL ‘red routes‘ manually – luckily it’s easy to search the data for their A road numbers. Here’s the City of London…


City of London

and London Borough of Bromley, often called the ‘greenest borough in London’…

London Borough of Bromley

Thick white lines symbolise streets measured in the study area and red lines as TfL red routes. Blue lines are borough boundaries. All that is left is to select all of the road categories we need to measure and use QGIS to calculate the sum, drop that back into the tree data spreadsheet and round-up the result to a whole number.

So, what’s the outcome?

  • 531 estimated street trees in City of London = 10 per 1km of suitable roads
  • 36,000 estimated street trees in London Borough of Bromley = 47 per 1km of suitable roads

47 trees per 1km in Bromley does sound like an over-estimate. That would mean a pair of trees either side of every suitable road at roughly 40m spacing! The GLA report does mention its concerns into quality of the estimates. Let’s hope the next release of the data is an improvement. In the meantime I need to automate the process a bit more and build my league table for the rest of the boroughs…