Category: Sustainability

Measuring Comparative Public Transport Accessibility for GB Cities

Posted in Mapping, Sustainability, transport, UncategorizedTagged , , , , , , , , , , Leave a comment

Although Greater London has an extensive transit network, this is not the case for many UK cities where underinvestment and privatisation has seen bus, metro and rail networks stagnate in recent decades, falling well behind European peers. Improving public transport is an important aspect of addressing the UK’s regional inequalities and poor productivity, and is a prominent issue for the 2024 general election.

Accessibility measures are an ideal tool to gauge the comprehensiveness and efficiency of public transport networks – they describe the ease with which populations can reach key services by different travel modes. The leading UK urban thinktank, the Centre for Cities (see their new Cities Outlook report 2024), has been doing some accessibility analysis of English cities compared to continental European cities, and this was recently republished in the Financial Times in an article on productivity challenges-

Centre for Cities report (left); Financial Time article (right).

It’s great to see accessibility analysis feature in the media. The measure used above however has some serious problems leading to nonsensical results (e.g. does Manchester really have half the accessibility of Liverpool and Newcastle?). The Centre for Cities measure uses a single time threshold (30 minutes) when we know that accessibility varies considerably at different time thresholds. It is based on a single destination point, when cities can have multiple employment centres. And it describes accessibility as a percentage of all city jobs, which means that the smaller the urban settlement is, the higher the accessibility result will be using this measure. In reality, larger city-regions have better jobs accessibility.

Creating Robust Public Transport Accessibility Measures – R5R and PTAI-2022
We can create much better and more reliable accessibility measures for UK cities. There have been significant recent advances. The open source R5R software has solved many of the computational challenges for accurately calculating public transport accessibility, allowing the calculation of full travel matrices for all possible trips and handling accessibility variation over time. In the UK, Rafael Verduzco and David McArthur at the Urban Big Data Centre have taken this one step further and pre-calculated accessibility indicators for all of Great Britain at a range of time thresholds in their Public Transport Accessibility Indicators dataset. This dataset is calculated using R5R, and is based on the median travel time across a three hour travel time window, 7am to 10am on a typical weekday (Tuesday 22nd November 2021), and uses the latest public transport service datasets such as the Bus Open Data Service. The results are at LSOA scale for GB only (no Northern Ireland), based on census 2011 zones (so I have used 2020 population data in the below analysis).

Origin and Destination Accessibility Measures
This article focuses on jobs accessibility, and this can be analysed from either the perspective of trip origins (residential-based accessibility to jobs) or from the perspective of trip destinations (workplace-based accessibility by residents). Both perspectives are complementary, and are developed below. For residential measures, if we take the average accessibility for all residents in a city then we get a good overview of how extensive and efficient the public transport network is. This requires city boundaries to define all the residents in each city. The analysis below uses the Primary Urban Area geography.

Public Transport Jobs Accessibility Trip Origin Results
The table and chart below show average accessibility to jobs for residents in all major GB cities by three travel time thresholds- 30 minutes, 45 minutes and 60 minutes. London’s accessibility results are inevitably much higher than any other GB city, being around 3 to 4 times higher at all three travel times, and emphasising just how big the gap is between the capital and all other GB cities. The 30 minute threshold describes shorter trips, and identifies higher density compact cities where residents are on average closer to employment centres. Small compact cities such as Cambridge and Oxford score well at 30mins (though note this is not the case at 45 or 60mins). Edinburgh and Glasgow have the highest residential average accessibility outside of London at both 30 and 45 minutes. This is due to Scottish cities historically following a higher density European urban model, and maintaining better public transport networks by avoiding some of the worst effects of privatisation.

The 60 minute accessibility measure picks up longer distance commuting on regional rail and metro networks. This is where the strengths of larger city regions such as Greater Manchester and the West Midlands are highlighted, with Manchester second and Birmingham forth in the ranking (Glasgow is third and also has a large regional rail network). Given their large populations, Manchester and Birmingham should however be scoring higher in absolute terms and closing the gap on London. Both have poor accessibility for the shorter 30 minute accessibility measure, reflecting the need for further inner-city densification (as the Centre for Cities have argued). For longer commutes, Manchester and Birmingham metro networks should also continue to be extended regionally. Leeds scores relatively well at 30 minutes due to its medium-density urban core, but it lacks a metro and is behind Birmingham, Glasgow and Manchester for the longer commuting times.

Peak Public Transport Accessibility by Trip Destination
We can also analyse accessibility by trip destination, which produces similar results to the trip origin residential measure but is more from the perspective of employment centres. The table below shows the peak accessibility by workplace within each Primary Urban Area, which is a measure of labour market size and agglomeration potential for the UK’s largest city centres. London retains its huge advantage with this measure, at 3 to 4 times higher than the next best cities. City-regions with larger rail and metro networks score better with the peak destination measure, with Birmingham and Manchester ranked second and third respectively, exceeding 2 million people at 60 minutes. Cities with strong rail connections to London, such as Reading and Crawley, also score highly at 60 minutes, but have much lower accessibility at 45 and 30 minutes. Smaller compact cities such as Edinburgh and Cambridge rank much lower by the destination measure compared to the residential analysis.

Both the trip origin residential average accessibility measure and the trip destination peak accessibility measure provide useful perspectives. The residential average measure is a good summary of the coverage and extent of public transport across a city, and how likely residents are to use public transport modes. The trip destination peak accessibility measures employment centre labour market size, and summarises the total number of people that can reach city centres by rail and metro. This is a better measure of agglomeration potential and is more closely correlated with city-region size.

Mapping the Accessibility Results
We can also map the results to view the geography of accessibility to jobs. Firstly the trip origin accessibility to jobs measure. This emphasises how large the area of high accessibility is across Greater London, with parts of Outer London and the South East having higher accessibility to jobs than residents in the city centres of the next largest cities, Manchester and Birmingham. The Primary Urban Area geography is also shown, which is the basis of the residential average accessibility chart and table shown above.

Next we map the trip destination accessibility to population measure. This has a very similar geography, but with more of an emphasis on city centres, as we are measuring average accessibility on a weekday 7am-10am when there will be more commuting services going to, rather than from, central areas. Again London has a huge advantage, peaking at 7 million people. We can also see the centres of Birmingham and Manchester reaching accessibility levels above 2 million people, while Glasgow, Leeds, Newcastle and Liverpool exceed 1 million.

Conclusion- Open Data and Software is Available to Create High Quality Accessibility Measures
With software such as R5R (see this workshop for an intro) and the exemplary and easy to use PTAI-2022 dataset from the UBDC, it is easier than ever to produce accurate public transport accessibility measures. The comparative accessibility analysis of GB cities shown here has highlighted the huge accessibility gap between London and all other UK cities. It has also shown the generally better accessibility performance of Glasgow and Edinburgh, and the high regional accessibility of Birmingham and Manchester which contrasts with their weaker accessibility in these regions for shorter travel times, which supports inner-city densification. There is no single perfect accessibility measure that answers all questions we are interested in – this analysis has confirmed that variation at different travel times reveals contrasting patterns in local and regional accessibility; that average and peak accessibility in cities emphasise different aspects of transit networks; and that trip origin and trip destination measures provide complementary perspectives. We therefore need to test a range of measures to understand accessibility patterns.

Future Improvements
This has been a relatively quick demonstration of the PTAI-2022 data and there are several areas for further improvements-

Including European cities for comparison would be very interesting, as the Centre for Cities explored in their original analysis. A recent major paper in Nature has shown how accurate international accessibility comparisons can be done- https://www.nature.com/articles/s42949-021-00020-2.

The PTAI-2022 dataset is a really good tool that makes GB accessibility analysis much more straightforward for researchers. Currently it uses the 2011 census boundaries, and the next update should use the 2021 boundaries allowing the latest census data to be used. Additionally, the current PTAI-2022 release uses 2021 public transport data, and updating this with the latest rail and bus data would also be a useful update. A related issue is that reliability on UK public transport networks can be poor, and that timetables can overestimate transit accessibility. This topic has been analysed by Tom Forth in this blog post.

This analysis has used the Primary Urban Area geography, which is a useful description of GB city-regions, but there are some issues with PUAs due to the underlying local authority geography. A few PUAs for medium-sized cities have quite large hinterlands (e.g. Sheffield) and this lowers the average accessibility measured in these PUAs due to lower accessibility outside of the urban core. A more thorough analysis of accessibility would need to test multiple urban geographies and gauge the extent of Modifiable Areal Unit Problem variation.

Can the Green Belt be Developed Sustainably to Ease London’s Housing Crisis?

Posted in City Planning, Housing, London, Mapping, Sustainability, transport, UncategorizedTagged , , , , , , 2 Comments

The housing crisis in London has become increasingly severe in the last decade with much higher prices, rents, and largely static incomes, while housing development volumes have remained consistently below targets. Green Belt reform is often cited as a solution to boost development, though this has been off the agenda during the last 13 years of Conservative government. Recent announcements by the Labour leadership, supporting Green Belt reform and setting ambitious targets for housing development, could change this state of affairs with the general election coming in 2024.

This article analyses housing development in the London region from 2011-2022 (full CASA Working Paper here), using the Energy Performance Certificate Data. There is strong evidence that the Green Belt is a major barrier to development and is in need of reform. On the other hand, there are very substantial challenges around the quality and sustainability of new build housing in the South East. The analysis shows that, outside of Greater London, new build housing typically has poor travel sustainability and energy efficiency outcomes. Any release of Green Belt land needs to be dependent on travel sustainability criteria and improved energy efficiency for new housing. Sustainable housing outcomes are much more likely to be achieved through prioritising development in existing towns and cities and in Outer London.

London’s Housing Affordability Crisis
House prices in London doubled between 2009 and 2016, pricing out households on moderate and low incomes from home ownership, and translating into rent increases, longer social housing waiting lists, increased overcrowding and homelessness (see Edwards, 2016; LHDG, 2021). Price rises are linked to on the one hand to the financialization of housing (exacerbated by record low interest rates and Help to Buy loans in the 2010s) and on the other a long period of low housing supply, stretching back to the 1980s and the erosion of public housing.

The impact is record levels of unaffordability, with Inner London average house prices reaching £580k and Outer London £420k in 2016 (see chart below). The median house price to income ratio for Inner London soared from 9.9 in 2008 to 15.1 in 2016; for Outer London the ratio increased from 8.2 in 2008 to 11.8. In addition to high prices, first-time buyers have also been hit with record mortgage deposit requirements, with average deposits reaching £148,000 for Greater London, compared to around £10,000 in the late 1990s (Greater London Authority, 2022). Owner occupation is now effectively impossible in Inner, and much of Outer, London for low and moderate income buyers.

There have also been substantial increases in prices across the London region. The map below shows prices per square metre in the South East showing four radial corridors of high prices extending beyond Greater London into the Green Belt. East London is increasingly mirroring West London with two radial corridors of higher prices extending north-east and south-east from Inner East London. These are the primary areas of gentrification in London in the last decade (discussed in previous blog post), squeezing out what was the largest area of affordable market housing. There is also a distinct spatial alignment between London’s Green Belt boundary and higher prices, which is evidence of regional housing market integration, and that Green Belt restrictions are pushing up prices.

New Build Housing Delivery in the London Region
Greater London has struggled to meet its housing targets in the last decade. The current London Plan target is for 52k annual completions, which, as can be seen in the graph below, London is significantly short of. The 52k annual target has been criticised as being too low, with other estimates of housing need calculating that 66k or even 90k houses per year are needed (LHDG, 2021). Given the extremely high prices, affordable housing tenures are needed more than ever, yet affordable housing delivery has fallen in the 2010s (although note there has been progress in affordable housing starts in the last two years). Finally, the recent impacts of the pandemic and high interest rates have hit market housing activity, meaning that London will very likely continue to miss its overall housing targets for the next 2-3 years.

We can look in more detail at the geography of housing delivery at local authority level in the scatterplot below. There is high development in most of Inner London, and some Outer London boroughs. These boroughs contain Opportunity Areas (major development sites in the London Plan): Canary Wharf in Tower Hamlets; the Olympic Park in Newham; Battersea Power Station in Wandsworth; Hendon-Colindale in Barnet; Wembley in Brent; Old Oak Common-Park Royal in Ealing; and Croydon town centre. Given that there are only a few Opportunity Areas in Outer London, this leads to relatively low delivery in most Outer London boroughs, and points to the need for a wider strategy for Outer London development.

Meanwhile, there is low development activity in nearly all Green Belt local authorities, much lower than London boroughs and also below the average for the rest of the South East. Green Belt restrictions affect both local authorities in the commuter belt and also Outer London boroughs as well (e.g. Enfield, Bromley) with 27% of Outer London consisting of Green Belt land. We can confirm how rigidly Green Belt restrictions are being applied using the official statistics, which calculate that the London region Green Belt land area was 5,160km2 in 2011 and 5,085km2 in 2022 (DLUHC, 2023). Therefore, only 74km2 or 1.4% of Green Belt land was released over the decade (this figure is for all development uses, not only housing), which is strong evidence of minimal change.

One final impact of the Green Belt can be seen by mapping development in the last decade as shown below. In addition to the patterns of high development in Opportunity Area sites, and generally low development in the Green Belt, there is a ring of high development activity just beyond the Green Belt boundary. This ring includes dispersed car-dependent development in semi-rural areas, and the expansion of medium-sized towns and cities such as Milton Keynes and Reading. This pattern looks very much like Green Belt restrictions are pushing development beyond the Green Belt boundary, creating sprawl-type patterns in several authorities. One important caveat is that several South East cities have strong economies in their own right, particularly technology industries in the Oxford-Milton Keynes-Cambridge arc, creating local development demands in addition to London-linked demand.

Potential for Green Belt Reform
With Greater London consistently falling short of housing targets, reform of the Green Belt has been cited as a promising solution (see for example Mace, 2017; Cheshire and Buyuklieva, 2019). The release of Green Belt land could greatly boost development and ease prices. Green Belt reform could also be a substantial source of revenue for austerity-hit local authorities, if authorities are given the powers to purchase Green Belt land at current use value and benefit from the land value uplift (this is part of the Labour proposals).

Traditional objections to Green Belt development focus on rural land preservation. Yet the Green Belt is massive in scale – 12.5% of all the land in England is Green Belt. London’s Green Belt is 5,085km2, or three times bigger than Greater London. Medium density housing development would take up a small proportion of this land. For example, building 100k dwellings at a gross density of 40 dwellings per hectare would add up to 25km2, or less than 0.5% of the London region’s Green Belt. Appropriate Green Belt reform could simultaneously allow for a moderate increase in development and improve environmental aspects of the Green Belt – the current environmental record of the Green Belt is mediocre on key measures such as biodiversity – through green infrastructure funding and principles of Net Biodiversity Gain. The land preservation arguments against Green Belt development do appear to be solvable. There are however further sustainability impacts from housing development to consider, including transportation and housing energy impacts, as discussed below.

Sustainability Impacts- Travel
Transport is the largest source of GHG emissions in the UK – 26% of all emissions in the latest 2021 data (DBEIS, 2023). The route to Net Zero requires both the electrification of transport systems and a significant mode shift from private cars to public transport, walking and cycling (HM Government, 2021). Greater London is a UK leader in sustainable travel, but this is not the case for the wider London region, much of which is car dependent. The analysis here uses car ownership and commuting mode choice data from the 2021 census to create a Travel Sustainability Index, as shown in the table below, which classifies Greater South East residents into 6 travel classes of around 4 million people. The South East covers a very wide range of travel behaviours, from an average of 20% commuting by car and 62% zero car households in the most sustainable class 1; to as high as 87% car commuting and 6% zero car households in the most car-dependent class 6.

Travel Sustainability Classes Average Statistics (2021 Census data)

Travel Sustainability ClassTravel Sustain. IndexCar
Commute %		Public Transport Commute %Walk & Cycle Commute %Car Owning Households %Residential Net Density (pp/km2)Total Pop. in South East
145-8220.348.526.438.351.5k3.56m
230-4541.633.220.961.532.1k4.03m
321-3060.618.117.674.725.0k4.03m
415-2171.610.914.283.320.2k4.16m
510-1580.06.510.989.416.4k4.34m
61-1087.33.66.794.111.1k4.29m

Mapping the travel sustainability classes highlights the stark travel behaviour differences between Greater London and the wider region. The Inner London population-weighted average travel sustainability score is 51.6 (class 1), and Outer London is 32.1 (class 2). The Green Belt is overwhelmingly in car dependent classes 4 and 5, with an overall population-weighted average of 16.4 (class 4). The Rest of the South East has a population-weighted average score nearly identical to the Green Belt at 16.5, emphasising the disappointing levels of car dependence in the Green Belt despite its rail infrastructure and proximity to London.

The patterns shown in the above map clearly present a challenge for Green Belt development, as new housing in the wider region risks extending patterns of car dependence. Car dependent areas include some locations next to rail stations (proximity to rail stations has been advocated as a criteria for prioritising Green Belt land for housing). We can directly measure the travel sustainability of housing development from the last ten years by matching the output areas locations of new housing to the Travel Sustainability Index scores. This is shown in the scatterplot below, where Inner London boroughs score highly with this measure, followed by Outer London. Much of the housing development in the wider region scores poorly in terms of travel sustainability, including in areas with high housing development such as Bedfordshire and Milton Keynes.

Although travel sustainability is generally low in the wider region, there are trends identifiable in the above results that can be used as basis for guiding more sustainable development. Several towns and cities show moderately sustainable travel outcomes, including the Green Belt towns Luton, Watford, Guildford and Southend, and wider South East towns and cities Brighton, Reading, Oxford, Cambridge, Portsmouth, Norwich and Southampton. Generally, development in existing towns and cities is likely to be more sustainable than developing smaller settlements and more dispersed rural areas. There are also noticeably better results in active travel-oriented cities such as Brighton and Cambridge. Overall, if we want Green Belt housing development to minimise travel sustainability impacts, then it would be most realistic to achieve this by extending existing towns and cities, both within the Green Belt and in the wider South East. Promoting development in Outer London boroughs also looks to be an efficient strategy given generally good travel sustainability levels in Outer London, and that Outer London is 27% Green Belt land.

Sustainability Impacts- Energy
Another important sustainability impact of new build is energy use and carbon emissions resulting from space and water heating, which we can estimate from the Energy Performance Certificate data as shown below. CO2 emissions per dwelling are considerably lower in Inner and Outer London, with overall London emissions per dwelling around two thirds of the value for the Green Belt and Rest of the South East. This is only partly due to smaller dwelling sizes, as CO2 emissions per square metre in London are significantly lower as well. The lower emissions in London housing can be explained by the much higher proportion of flats and also the use of community/district heating, with three quarters of all new build in Inner London and 47% of new build in Outer London connected to community heating networks. The community heating approach is only efficient for high density developments. For medium and lower density developments, air and ground source heat pump technologies are a key technology for improving energy efficiency and replacing gas boilers. The statistics from 2011-22 are very disappointing on this front, at 4% of new build with heat pumps in the Green Belt and 6% in the Wider South East.

New Build Annual Average CO2 Emissions and Energy Summary 2011-2022 (Data: EPC 2023)

SubregionCO2 per Dwelling
(tonnes)		CO2 per m2 (kg)Energy Consumption
(kWh/m2)		Community
Heating %		Heat Pump % (air + ground)
Inner London0.9312.972.975.22.7
Outer London1.0415.387.246.92.8
Green Belt1.6018.7106.97.93.5
Rest of South East1.5317.297.75.75.9
All Subregions1.3416.392.527.04.3

The average annual CO2 emissions by dwelling are summarised at the local authority level in Figure 19 (note y axis starts at 0.5). Similar to the travel sustainability results, London boroughs have considerably more sustainable results. Town centres in the South East again are the best performing outside of London, including Cambridge, Southampton, Eastleigh, Reading, Luton, Watford, Woking and Dartford. As the chart shows average CO2 per dwelling, there is a connection between affluence and dwelling size, with higher income boroughs such as Richmond Upon Thames and particularly Kensington and Chelsea, having high emissions. Overall however, energy efficiency is much better in London boroughs and this is a further challenge for the sustainability of Green Belt development. Similar to the travel sustainability analysis, the results point to the extension of existing towns and cities, and Outer London development, as the most sustainable development strategies.

Summary
There is a widespread consensus that London needs to build more housing to meet demand and try to reduce record levels of unaffordability. Yet London has been consistently short of meeting housing targets for the last decade, despite substantial growth in Inner London. Green Belt restrictions do appear to have played a major role in constraining development, with low levels of new build in Green Belt local authorities, and in Outer London boroughs with extensive Green Belt land. There is also a significant price premium in Green Belt areas compared to the wider South East.

This analysis agrees with research advocating Green Belt reform. Travel sustainability conditions are needed to avoid this reform producing highly car dependent housing, such as has been occurring in Central Bedfordshire and Milton Keynes (where the East-West should have been built much earlier). Pedestrian access to rail stations is a sensible starting point for prioritising Green Belt land for housing, but it is not sufficient to produce sustainable travel outcomes in the Green Belt. The aim should be for new housing to have local access to a range of services (e.g. retail, schools), providing sustainable travel options for multiple trip types. Another related issue is the need for more sustainable energy efficiency measures in medium density new build housing. There is little evidence in the EPC data for adoption of key housing technologies such as heat-pumps and solar PV. Widespread adoption of these technologies is needed for sustainable development at scale in the Green Belt. Other studies have also identified poor design and planning in new build housing in the UK (see Carmona et al., 2020), and this needs to change as part of any plan to increase the volume of new housing.

Green Belt reform would have to come from national government, changing the very restrictive current National Planning Policy Framework to allow authorities with housing shortages to develop Green Belt land of low environmental quality near services, and to use land value uplift to fund services and affordable housing. It would be logical to give powers to the GLA (and other combined authorities) for the strategic coordination of this development within their boundaries, given the GLA’s strong track record on sustainable housing delivery. It is difficult however to envisage large scale change happening in the South East without national government also organising improved regional coordination and planning. This analysis identifies better travel sustainability outcomes for new build in larger towns and cities in the South East, and supports the urban extension model for development in the Green Belt. There are many candidate towns in London’s Green Belt for urban extensions, including Luton, Guildford, Watford, Maidenhead, Hemel Hempstead, Chelmsford, Basildon, Reigate and Harlow. This larger scale solution is politically more challenging, and would again require leadership and coordination from national government.

—————————–

Read the full CASA Working Paper.

This research is part of the ESRC / JPI Europe SIMETRI Project.

World Population Density Map Update with GHSL 2023

Posted in City Planning, Mapping, Sustainability, transport, VisualisationTagged , , , , , Leave a comment

The European Commission JRC recently released a new 2023 update of the Global Human Settlement Layer (GHSL) data. This update has greatly improved the GHSL data, with a 10 metre scale built-up area dataset of the entire globe which has been used to create a 100 metre scale global population density layer. The level of detail for cities and rural areas is impressive, and it overcomes the limitations of previous releases of the GHSL. I have updated the World Population Density Map website to include this new 2023 data, with both the cartography and statistical analysis now based on the new data.

Improved Level of Detail for Cities and Rural Landscapes
The new GHSL 2023 data has produced a much more detailed 10 metre dataset of built-up area (using recent European Space Agency Sentinel data), and this is the basis for creating the updated population layer. The results are much improved, particularly for complex rural and peri-urban landscapes in the Global South, such as for India shown below. The tens of thousands of small villages are identified and used to more accurately distribute India’s huge population. This is also the case for other key regions such as Sub-Saharan Africa, and China.

The added level of detail also improves the representation of cities, with more accurate density analysis, and improved techniques to differentiate residential from industrial and commercial urban land uses. Previous releases of the GHSL were underestimating urban densities for cities where census data was weaker, but this appears to no longer be the case. The dataset can now be used for more accurate comparisons of population and density for cities across the globe. Example images for Shanghai and New York City are shown below.

Country Density Profiles – the Diversity of Human Settlement
The statistical analysis on the World Population Density Map website has also been updated using the 2023 GHSL data, so you can view the density profiles for all countries around the globe. Some highlights are shown below.

To complement the graph of the population in each density category, this updated version of the World Population Density Map includes Population Weighted Density statistics for each country and city. Population Weighted Density is a measure of the typical density experienced by residents in the country/city, in this case using the 1km2 scale GHSL data. The PWD is calculated by weighting each 1km2 cell according to the population, summing all the cells for the city/region, and then dividing the sum by the total population of the country/city (i.e. the arithmetic mean). This is a more representative measure than standard population density, which is affected by low density suburban/peri-urban and rural land, even where the population in these areas is relatively low.

China and India have very high density cities, but their large rural populations translate into moderate Population Weighted Density statistics overall. India is 9.9k pp/km2 and China is 8.9k pp/km2. The table below shows the top 20 countries by Population Weighted Density using the 2020 data-

Rank (by PWD 2020)Country NamePopulation Weighted Density 2020 (pp/km2)
1Singapore30.9k
2Republic of Congo25.2k
3Somalia24.1k
4Egypt21.8k
5Comoros17.4k
6Djibouti17.2k
7Iran16.8k
8Yemen16.7k
9Jordan15.8k
10North Korea14.8k
11Democratic Republic of the Congo14.2k
12Bahrain13.9k
13Colombia13.5k
14Equatorial Guinea13.5k
15Turkey13.5k
16Morocco13.4k
17Bangladesh13.3k
18Taiwan12.9k
19South Korea12.7k
20Western Sahara12.6k

For comparison, the equivalent Population Weighted Density figure for the UK is 4.1k, France is 3.7k and Germany is considerably lower at 2.7k. The USA is renowned for its low density living and suburban sprawl, and the Population Weighted Density measure for 2020 is 2.2k. This is the lowest figure for any large developed country in the world. Smaller developed countries have similar figures to the USA, including New Zealand, Norway and the Republic of Ireland.

Analysing the World’s Largest City-Regions Using the GHSL
The Built-Up Area and Population layers in the GHSL are used to define a settlement model (GHSL-SMOD) layer, which classifies land into urban and rural typologies. We can use this layer to define the boundaries of city-regions across the globe. This has been done using continuous areas of the highest urban category (urban centres) for the 2020 data. When you hover over cities on the World Population Density website, these city boundaries are highlighted-

This land use based method of defining city-regions produces different estimates of city populations to analyses based on administrative boundaries. The GHSL method generally emphasises large continuous urban regions, such as the megacity region of the ‘Greater Bay Area’ in China shown above, which has formed from the fusion of Guangzhou, Shenzhen, Dongguan and Jiangmen. This is the largest city-region in the world by this measure, with a population of 43.8m in 2020 (rapidly developing from a base of 5.8m in 1980). The top twenty city-regions in the world are shown below-

Rank (by Pop. 2020)City-Region NamePopulation 1980Population 2000Population 2020Pop. Weighted Density 2020 (pp/km2)
1Guangzhou-Shenzhen-Dongguan-Jiangmen5.8m30.9m43.8m20k
2Jakarta16.1m26.3m38.7m13.4k
3Tokyo27m31.3m34.1m10.2k
4Delhi8.3m19.1m30.3m29k
5Shanghai6.7m15.1m27.8m27.9k
6Dhaka6.2m14.8m26.8m47.9k
7Kolkata16.3m22.9m26.7m36.4k
8Manila11.3m18.3m24.8m27.1k
9Cairo9.8m16.6m24.5m44.9k
10Mumbai11.3m18.4m22.9m52.4k
11Seoul13.3m19.7m22.7m19.8k
12São Paulo13.7m17.4m19.7m14.5k
13Beijing7.2m11.6m19m20.2k
14Karachi5.8m10.9m18.7m48.8k
15Mexico City13.8m17.9m17.8m13.2k
16Bangkok5.4m9.2m17.4m11.6k
17Osaka17.2m16.8m15.6m8.1k
18Moscow9.9m11.9m14.9m16.7k
19Los Angeles10m13.1m14.5m4.6k
20Istanbul6.1m10.4m14.3m25.2k

One of the most impressive aspects of the GHSL is that it is a timeseries dataset going back to 1975. Therefore we can create historical indicators such as the population change data shown in the table above. Many cities have more than doubled, or even tripled in population size since 1980, including Delhi, Shanghai, Dhaka and Karachi. Rates of growth in the USA, Japan and Europe are inevitably much lower, as seen in Tokyo and Los Angeles in the table above. Tokyo is often measured as the world’s largest city (for example in the UN World Urbanization Prospects), though with the GHSL method Tokyo the third largest at 34.1m in 2020. Tokyo is also distinctive in terms of its Population Weighted Density at 10.2k pp/km2. While this figure is more than double the density of Los Angeles, Tokyo’s medium density is much lower than cities in China and South Asia. Incredibly, Mumbai’s density figure is five times higher than Tokyo at 52.4k pp/km2, and Karachi is not far behind at 48.8k.

Credits and Links
The Global Human Settlement Layer is published by the European Commission Joint Research Centre. All the GHSL layers are available as open data and can be downloaded on their website. The GHSL population data uses as an input the Gridded Population of the World data by CIESEN at Columbia University.

The World Population Density Map website has been created by Duncan A. Smith at CASA UCL. The mapping data is hosted on a tileserver at CASA UCL kindly set up by Steven Gray. The interactive mapping layers are hosted on Carto.

Post-Pandemic Changes in Travel Behaviour: Evidence from the National Travel Survey England 2021

Posted in City Planning, Sustainability, transportTagged , , Leave a comment

The pandemic and subsequent lockdowns have seen the largest and most sustained disruptions to travel behaviour in most of our lifetimes. Stay-at-home policies have fuelled a dramatic increase in remote working, and wider online substitution of other activities such as shopping and socialising. In sustainability terms, the pandemic has severely hit public transport and incentivised car travel, but has also likely reduced travel distances overall as well as encouraging new patterns in active travel. The big question is to what extent pandemic related changes are turning into longer term shifts in travel behaviour patterns.

This post looks at timeseries travel data across the last three years, and then summarises results from the recently published National Travel Survey data 2021 for England, with a particular focus on trip purpose and differences between London and England as a whole.

Transport Use Timeseries Data from DfT
The Department for Transport have continually updated a very useful timeseries on how busy different transport modes have been in England throughout the pandemic. This index integrates many different datasets and is intended as a broad summary of trends (see methodology here). The graph below summarises this data, which is indexed to February 2020. The overall picture is of huge disruption in 2020, continued disruption with a transition towards recovery in 2021, and then what looks like settling into a new normal in 2022.

The chart paints a mixed picture in sustainability terms. Car travel has been the fastest transport mode to recover after each of the national lockdowns, and was back to near normal levels as early as summer 2021. While this is a challenge going forward, it could potentially have been worse. The pandemic could have resulted in substantial increases in car travel. Instead, there is a minor reduction to about 96% car use in the DfT data, sustained into 2022 (in per-capita terms this reduction will be more substantial given population increases). Note the motorised vehicle index that includes freight reaches 100% of pre-pandemic levels in 2022, possibly due to more online delivery traffic.

Public transport has been much slower to recover, falling to less than 50% of passenger numbers in 2020, increasing substantially throughout 2021 and then settling around 70-85% of pre-pandemic passenger numbers in 2022. Rail and tube travel were hardest hit in 2020 due to the widespread fall in commuting and these modes have taken longer to recover than bus travel. It is difficult to gauge whether public transport levels have now levelled off around the 75% level, or will continue to recover further in 2023 (the rail and tube strikes in summer 2022 may have curtailed further increases).

A positive sustainability story comes from the cycling data from the DfT. This is a less reliable metric, but nonetheless indicates growth in active travel, albeit from a low base in 2019. The annual variation in cycling in the DfT data between 2020 and 2022 is interesting. The initial 2020 increase in cycling makes sense, as there was a big growth in active travel for households locked down in their local area. This falls to 2019 levels in 2021, and then rebounds in 2022. Perhaps the fall off in new cyclists has given way to more practical longer term adoption of cycling in 2022.

National Travel Survey Data 2021
The results for the National Travel Survey 2021 data were released at the end of August 2022. This long running survey records full travel diaries of thousands of residents in England, allowing analysis of topics such as trip purposes, walking trips and demographic analysis. There have been some data gathering challenges during the pandemic (see data quality report). The overall picture from the DfT chart below is that the 2021 NTS data is very similar to the 2020 data in terms of per person trips and annual distance recorded. This suggests that the NTS 2021 is not picking up much of the 2021 travel recovery that is shown in the DfT timeseries data we discussed above.

We can break down the annual trip distance per person by transport mode. The chart below compares the years 2019, 2020 and 2021. The results backs up the conclusion that, at the national level, the 2021 data is very similar to 2020. There is big reduction in car miles of 30%, while public transport levels are at around half the levels of 2019. There is a small increase in walking and cycling distances, though this falls back to 2019 levels in 2021.

Interestingly, the equivalent chart for London shows a very different picture in terms of travel behaviour responses. Car use increased marginally in 2020 (probably substituting for public transport trips) then falls in 2021, though this reduction is lower than the national picture. Bear in mind London mileages are around a third of the national average, so there may be fewer optional trips to cut. Meanwhile, public transport records a significant recovery in 2021 for rail and underground, much closer to the DfT time series analysis shown above (note the bus recovery is lower than expected). Walking and cycling follows the national picture by increasing in 2020 and then falling in 2021, though note that the 2021 cycling level is substantially up on the 2019 figure.

Overall the broad 2021 National Travel Survey results are fairly close to the 2020 results at the national level. In comparison to the DfT Transport use timeseries, it appears to be closer to the 2020 picture rather than the 2022 trend of a substantial recovery in transportation use. So we may have to wait for the National Travel Survey 2022 to confirm which changes are really sticking in terms of longer term behaviour. The London data is interesting, as it much more clearly shows a recovery in public transport travel in 2021, and a different picture for changes in car mileage, as well as a substantial increase in cycling.

Trip Purpose Analysis
An important question is the type of trips most affected by the pandemic, and whether these changes are becoming longer term trends. The charts below show the trips per person per year and total distance per year between 2019 and 2021. As expected commuting is taking the biggest hit in terms of trips and distance, falling by 36% in distance terms and with only small signs of recovery in the 2021 NTS data. Drops in business travel are even larger, more than halving. Interestingly shopping trips have taken nearly as big a hit as commuting, with distances falling 26%. There has been a widespread trend towards online supermarket deliveries and online shopping more generally post-pandemic and it looks like this behaviour has continued into 2021. The 2021 NTS even shows shopping trips and distances falling again in 2021 from the 2020 level. Alongside commuting changes, shopping travel behaviour looks to be the major trip type that has been cut, possibly for the longer term.

Outside of commuting and shopping, other trip types with big reductions include holidays, business, and entertainment. In contrast day trips increased and walking trips nearly doubled (though both fell back marginally in 2021 from the 2020 peak). Visiting friends at their home also continued during the pandemic, with a more minor reduction in trips and distances.

Finally we repeat the distance trip purpose chart for London. Commuting takes an even bigger hit in London, falling by 48% in 2020, then moderately picking up in 2021. Business trips fell by a huge 67% and there is little sign of recovery. In contrast some trip types that declined in 2020 are nearly back at 2019 levels, such as education and education/other escort trips. The trips with the biggest increases in 2020, visiting friends at private homes and day trips, have also returned to their 2019 levels in the 2021 data. Walking trips have however remained considerably above their 2019 level, indicating that the active travel increase is looking more stable for London.

Summary

The DfT timeseries data shows travel patterns settling into a ‘new normal’ after more than two years of disruption. Car travel is only marginally down on pre-pandemic levels, while public transport is around 70-85% of the passenger numbers from 2019. There are some encouraging signs for active travel after increases in leisure walking and cycling trips, though the situation is dynamic.

The National Travel Survey 2021 data records a similar picture overall to 2020 in terms of major disruption- distances and trips down substantially. Car miles are down 30%, though the DfT timeseries data suggests this will not be the case in the 2022 data. Public transport remains around half of 2019 levels.

The trip types with the biggest reductions are commuting and business as expected, but also shopping (and the related travel type of personal business). Shopping trips fell again in 2021, and this looks, like commuting, to be a case of significant online substitution, possibly long term.

The London NTS data is very interesting, with some significant differences. Car use appears to have marginally increased in London in 2020, before falling in 2021 below 2019 levels. Public transport has staged a major comeback in London in the 2021 data, and there is also a sustained increase in cycling. The NTS data is also disaggregated into urban/rural categories and it would be interesting to test whether these trends are London specific or shared with other major English cities.


A Compact City for the Wealthy? Continuing Inner London Gentrification and Impacts on Accessibility Inequalities

Posted in City Planning, London, Mapping, SustainabilityTagged , , , , , , Leave a comment

We have a new paper out in the Journal of Transport Geography- “A compact city for the wealthy? Employment accessibility inequalities between occupational classes in the London metropolitan region 2011“. The paper explores how the increasingly affluent nature of Inner London has improved sustainable travel opportunities for more affluent professional and management classes, while less affluent groups have increasingly been priced out to lower accessibility Outer London locations.

The Continuing Gentrification of Inner London
The gentrification of Inner London was first recorded by Ruth Glass back in the 1960s, with middle class residents moving into largely working class neighbourhoods as London’s economy began its long evolution from manufacturing towards service jobs. This process has continued for decades, ultimately transforming most of Inner London. In the 21st century, some researchers have argued that gentrification has stalled (perhaps because there are few neighbourhoods left to gentrify) or has entered a different phase (e.g. processes such as super-gentrification and new-build gentrification as discussed by Davidson and Lees).

This research uses the Standard Occupational Class data as the basis of measuring social class. This classification was found to correspond to differences in income, as well as to a distinct residential geography. In particular, the three most affluent groups (Management, Professional and Associate Professional) cluster together, resulting in the social geography we can see in the map below using the 2011 Census data. There is a clear clustering of professional classes in Inner-West London, with two prominent radial corridors extending northwards through Camden, Islington and Hampstead; and south-westwards through Kensington, Wandsworth and Richmond. Concentrations of non-professional groups are mainly in Outer London to the east, north-east and west, with only smaller pockets remaining in Inner London. This analysis largely matches the description of Inner London now being dominated by professional classes, with lower income groups increasingly in Outer London (with some exceptions remaining in Inner East and South-East London).

Professional Classes (Manag., Prof. & Assoc. Prof.) Residential Percentage 2011. Data: Census 2011 (Office for National Statistics, 2016).

As well as mapping the 2011 geography of occupational class, we looked at more recent changes to see if gentrification is continuing or has slowed, using the ONS Annual Population Survey. Analysing changes between 2006 and 2016, we found had substantial gentrification had continued in Inner London, as shown in the table below. The Management, Professional and Associate Professional groups all grew as a proportion of the Inner London population, while all other occupational classes fell proportionally (green cells are above the average for the metro region, and orange cells are below the average). Interestingly, the biggest growth was in the Management and Professional classes, rather than the younger Associate Professional class, arguably more in line with super-gentrification processes. In contrast, there are proportional increases in several lower income classes in Outer London.

Sub-Regional Occupational Class Percentage Point Change by Residence 2006–2016 (final 2016 sub-regional percentages in brackets)

We can also explore these changes at the more detailed level of local authorities, and show that even more dramatic changes are occurring at the local level. In the chart below, each Local Authority is shown as a trajectory connecting its position in 2006 to its position in 2016 in relation to the percentage of professional classes and the total working population. Generally, Inner Greater London Authority (GLA) boroughs experience high working population growth combined with large increases in the proportion of professional classes. Boroughs with a long history of gentrification, such as Camden and Islington, are higher up in the chart reaching 70% professional classes, while more recent gentrifiers, such as Lewisham and Southwark, are rapidly gentrifying from a lower base. Outer GLA boroughs also show substantial population growth but with lower levels of change in professional classes, and decline in some cases. The exceptions are mainly in South and South-West London, with Croydon, Sutton and Richmond all gentrifying. Outer Metropolitan Area (OMA) local authorities have a mixed picture, with some increases in professional occupational classes with minimal working population growth; while some lower income towns such as Luton and Harlow are not gentrifying.

Local Authority Trajectories for Combined Percentage of Professional Occupational Classes and Total Working Population for 2006 and 2016. Data Source: Annual Population Survey 2005–2017.

What Impacts Does Inner London Gentrification have on Accessibility to Jobs?
We would expect that the dominance of more affluent classes in Inner London translates into accessibility advantages for these classes, as Inner London has substantially better accessibility opportunities by public transport, walking and cycling. We were particularly interested in accessibility by more affordable travel modes in this research. Bus travel is in general considerably cheaper than other public transport options in London. This is reflected in higher rates of more affordable bus and walking trips by lower income classes in the 2011 Census data. We can see in the table below that the three lowest income classes (6, 7 & 9) have around three times higher rates of bus travel and two times higher rates of walking than the most affluent three classes (1, 2 & 3)-

We used network analysis to analyse accessibility differences (see working paper on accessibility model). The analysis was carried out using the 2011 census data. The box plot below shows the cumulative accessibility to jobs for 60 minutes travel by Car, Public Transport (all modes) and Bus Only for the occupational classes. We can see differences between classes, particularly for public transport and bus trips, though there is also much variation within each class.

GLA 60mins Cumulative Accessibility to Employment by SOC Groups: Absolute Results

The accessibility differences between occupational classes can be more clearly seen by plotting differences between how the average accessibility for each group varies from the average accessibility for the entire working population, as shown below. Note in this chart the accessibility differences are normalised by travel mode, so the differences between travel modes in the chart above are normalised in the chart below. We can see clear consistent accessibility advantages for the top three occupational classes, particularly for more affordable slower modes- walking and cycling. The remaining occupational classes have below average accessibility to jobs, particularly for the more car oriented Skilled Trades and Process groups.

Greater London Authority 60mins Cumulative Accessibility to Employment by SOC Groups: Relative Differences in Occupational Class Mode Means and Mode Means for Total Population

The results for bus and walking modes is a particular accessibility challenge. Accessibility by these more affordable modes is generally low in absolute terms outside of Inner London. For the bus mode, less than half the number of jobs are reachable at typical commute times compared to the full public transport network. Given that lower income groups are the most frequent bus and walking commuters, and that these classes are increasingly being priced out of Inner London, these limitations are a significant accessibility challenge going forward.

What Policies Can Planner follow to Mitigate this?
In terms of transport policy, this research supports efforts to improve the affordability and connectivity of public transport for lower income populations. This is indeed a priority of the current London Mayor Sadiq Khan, who has committed to freezing public transport fares, and has reformed ticketing to allow multiple bus journeys on a single fare. These measures help offset travel costs for lower income residents in Outer London.

The main policy conclusion is the importance of housing policy in influencing accessibility outcomes in the study area. Low and moderate income groups are being priced out of public transport accessible areas. Without a step-change in the delivery of genuinely affordable housing in accessible locations, the increasing dominance of Inner London by professional classes will continue, resulting in greater accessibility inequalities, and likely increased travel costs for lower income classes.

Note on Covid-19 and Travel Inequalities
This research was completed in 2019, before the recent COVID-19 pandemic. The pandemic has in the short term shut down public transport networks, and greatly disadvantaged millions of city residents around the world. Longer term it is possible that the pandemic will reduce the attraction of inner city areas such as Inner London, due to perceived risk of future pandemics as well as the acceleration of telecommuting and home-working trends. The overall effect could be to slow gentrification processes, although this is difficult to predict. The alternative view is that  London will recover and adapt as it has done following many crises in the past. East Asian metropolises offer a good model of how to built resilience following their response to the earlier SARS and MERS outbreaks.

The wider economic impacts are clearly also important. Certainly we are in line for a very large recession, hitting important sectors such as tourism and hospitality. More specifically in London, the recession may hit development viability for affordable housing, and is a real headache for public transport operators. Transport for London was in financial trouble before the crisis, and is currently dependent on government bailouts to keep running. This will likely curtail the ability of the Mayor to maintain lower public transport fares, and so impact the kind of transport accessibility inequalities this paper discusses.

 

 

 

Planning a Cycling Revolution for Post-Lockdown London

Posted in City Planning, Cycling, London, SustainabilityTagged , , , Leave a comment

Overview
London and the UK as a whole have been severely affected by the Covid-19 pandemic, creating multiple health, social and economic crises. Social distancing requirements have drastically reduced the safe capacity of public transport services. The London Mayor and Transport for London (TfL) have quickly responded with a radical plan for transforming Inner London into a walking and cycling city, with major streetspace reallocation and restrictions on car use. Where should this new active travel infrastructure be prioritised? Will it substantially change travel behaviour for the better? This article looks at plans for new cycling infrastructure in Inner London, and includes recent UCL research by Nicolas Palominos.

The Green Active Travel Response to the Covid-19 Pandemic
The concentration of people, activity and global connections in cities is normally their greatest asset, yet now is the source of acute risk and vulnerability during the coronavirus pandemic. Rail, metro and bus systems enable millions of daily trips to be made efficiently and sustainably, but their crowded nature is incompatible with social distancing. Recent estimates of how many passengers can safely use buses, trains and metro systems in London while enforcing a 2 metre social distancing rule are around 15% of full capacity. This restriction is devastating for transit cities to function, preventing people getting to work, school and basic services; as well as crippling the fares revenue for public transport providers. Although there are plausible scenarios where the 15% limit could be moderately relaxed (such as through mandatory mask-wearing and temperature screening, or an adjustment of the UK social distancing rules to 1.5m/1m) we are effectively looking at months (in the worst case years) with public transport operating at a severely reduced capacity.

With public transport now hampered for the short-to-medium term, there is a real danger that there will be a big increase in car use as lockdown restrictions are eased, undoing much of the recent progress towards sustainable cities and tackling urban problems of congestion, poor air quality, carbon emissions and obesity. The alternative is to dramatically increase walking and cycling through a major reallocation of street space to these active travel modes. Measures for new cycle and pedestrian lanes have recently been proposed by many cities including Milan, Paris, Brussels and Bogota, and are being widely discussed in many more. The aim is to provide safe and healthy travel alternatives to increased car use, and to enable cities to continue to function with public transport restrictions. London Mayor Sadiq Khan has joined these progressive cities by advocating an active travel transformation for London with the Streetspace for London plan.

London’s Fragmented Cycle Routes
Cities that have achieved very high rates of cycling, such as Amsterdam and Copenhagen, have done so with comprehensive networks of segregated cycle lanes, allowing all residents to cycle safely, not just experienced cyclists. London has been struggling to catch up with these cities, and still has a fragmented and incomplete network of cycle lanes. The map below shows Inner London cycle lane data in 2019. The dark blue lines are physically segregated cycle lanes (i.e. a physical barrier between the cycle lane and the road). These are mainly found on CS2 (Cycle Superhighway 2) in East London, CS3 (the only complete East-West route through Central London), and parts of CW6 (Cycleway 6) which is the main north-south link across Blackfriars Bridge that has some sections in Central London that are still under construction.

Outside of these segregated routes there is a scattered patchwork of cycle lanes that are based on road markings without a physical barrier with the carriageway, shown in light blue on the map. There are also bus lanes on many major roads (dark red on the map) which offer some protection for cyclists, but are a long way from the safety of fully segregated lanes. Overall Inner London’s cycle network has many gaps that need filled, particularly in Central, West, North-West, North-East and South-East London. This infrastructure gap sits in sharp contrast to the aim of drastically increasing cycling during the post-lockdown period. The lack of segregated routes increases the risk of cycle collisions with cars, and will discourage public transport users from making the switch to cycling.

The Proposals from the Mayor and TfL
In the Streetspace for London plan, the Mayor has proposed developing a network of new walking and cycling routes to allow Londoners to walk and cycle while social distancing, and to redirect public transport trips to active travel rather than driving. Cycling will be critical for public transport substitution, as most tube and bus journeys are 4-10km, considerably longer than typical walking distances (discussed more below). Given the fragmented nature of current cycle routes, major changes and interventions are required to achieve these aims.

The highly ambitious nature of the proposals were made clear last week with a major plan for new Car Free routes to be introduced in the City of London (below). This plan would transform many of London’s busiest roads into bus, cycling and walking corridors, radically improving cycle access across Central London, particularly to major rail stations. It includes:

  • Completing Cycleway 6 with a car free Farringdon Road leading to Kings Cross Station;
  • A car free Waterloo Bridge, leading to a whole new car free north-south route via Kingsway and Southampton Row to Euston Station;
  • A car free London Bridge, leading to two segregated routes, the first via Bank to Moorgate, and and the second via Bisohopsgate to Liverpool Street station;
  • Upgraded east-west routes, with a car free Theobalds Road to Old Street; and improved cycle provision on High Holborn and Cheapside (could this be extended west to Oxford Street, and east to meet Cycleway 2 at Aldgate?);

These proposals show a comprehensive commitment to the active travel plan, and include major restrictions on driving to make this happen. In addition to the Car Free routes, the Congestion Charge and low emission zones have been reinstated, and the cost of the Congestion Charge will rise to £15, and hours of operation extended to 7am-10pm 7 days a week.

The Central London plan also shows that the City of London (the borough authority that controls London’s financial district) share the Mayor’s vision, as this proposal can only happen with their close cooperation. Indeed the City of London had already developed a longer term plan for removing cars from most of the City, and this proposal fits closely with their vision.

While Central London is vital for the Mayor’s active travel plans, the vast majority of journeys begin outside the City. Clearly a much wider vision for all of London is needed. TfL have stated they will create 30km of new permanent cycle lanes this summer. They have released a London-wide map of proposed new ‘Streetspace Routes’ on their website (below). While this map is very much in development (it does not yet have the City of London routes above added), there are several very significant proposals, including-

  • Segregated cycle lanes on Euston Road- a vital busy highway connecting major rail stations (Paddington, Euston, Kings Cross), as well as linking to the north-south routes in the above Central London map. The Euston Road proposal extends all the way west to Acton;
  • adding a further set of new routes in West London, including Cycleway 9 from Hounslow to Brentford and Kensington Olympia;
  • speeding up the completion of CW4 from London Bridge to Greenwich and Woolwich in South East London;
  • improving CS7 to Brixton and Clapham in South West London;
  • completing a North-West link from Regents Park to Golders Green;
  • adding a new link from Hackney to Canary Wharf;

Predicting Where Cycling Demand is Likely to be Highest
While there are lots of excellent proposals in the TfL map above, it does not yet form a complete network (it also does not differentiate between the quality of existing cycle routes which are highly variable). We can use network analysis to consider where cycling demand is likely to be highest, and where new infrastructure should be prioritised. The analysis below by Nicolas Palominos at CASA is based on shortest paths between underground and rail stations, and is indicative of where active travel trips are most likely to be made if passengers are prevented from taking the tube/bus/train (see full working paper here). The map highlights the core network of Inner London’s most prominent links, including its oldest Roman Roads- Edgware Road, Kingsway, Oxford Street. Promisingly, there is considerable overlap between these critical shortest path routes and the cycle network proposals by TfL. We can see the importance of the new north-south Waterloo Bridge-Southampton Row link for example; Euston Road’s importance is clearly shown; as is Theobald’s Road to Old Street. TfL have completed similar analysis in their Strategic Cycling Analysis from 2017, and we can assume this is guiding their plans.

There are however some important links missing highlighted through the network analysis. The most significant missing link is London’s most important east-west route that runs along Hyde Park through Oxford Street all the way to Cheapside and the City (following the route of the Central Line). The City of London proposals cover the eastern section of this vital route, from High Holborn to Bank. There needs to be similar cycling upgrades for the western section: Oxford Street, Bayswater Road and Holland Park Avenue. This intervention requires the Mayor to cooperate closely with the City of Westminster authority. Politically this is difficult, as Westminster has been blocking high profile proposals from the mayor in recent years, most notably the pedestrianisation of Oxford Street. Coming to some agreement will be essential for the western expansion of the cycle network to be joined together effectively.

Other key routes that are not yet part of the proposals include Edgware Road; Peckham Road; Essex Road; Old Kent Road; City Road; Commercial Street; Kingsland Road; and Seven Sisters Road amongst others. These are discussed further in the working paper. Edgware Road is currently the main north-south driving route that avoids the Congestion Charge, and giving space to cycling would be difficult to implement here (although Park Lane is included in TfL’s plans, and indeed the Park Lane cycle lane is already installed on the western carriageway adjacent to Hyde park). The Swiss Cottage to Golders Green link looks like the proposed north-west alternative to Edgware Road, and would benefit from completion through Regents Park to the Euston Road link. The network analysis strongly agrees with the Euston Road cycle provision, though in this analysis this critical link continues further to Pentonville Road, City Road and Commercial Street encircling the City.

Overall, the TfL Streetspace proposals look very promising, and need to be further linked together and expanded as discussions with more boroughs continue. This will also need much more emphasis on Outer London centres and local shopping and school routes, important topics which are not covered in this article.

Estimating the Number of Cycling and Walking Trips Post-Lockdown
What might the the number of trips by different travel modes in London look like in the coming months? This is highly uncertain. Some modelling figures from Transport for London quoted in press releases are anticipating radical changes in behaviour, up to a ten times increase in cycling trips, and up to a five times increase in walking trips. Are these figures realistic?

The chart below shows the distribution of trips pre-lockdown (from TfL Travel in London Report 12), and a series of post-lockdown guesstimate scenarios to discuss where we might be heading. Normally in London we are looking at around 27 million trips per day, with 9.5m by transit, 10m by car, 7m by walking and 0.7m by cycling. The lock-down figures in the chart below are based on TfL’s statement that underground passengers have been at 5% of normal levels, and car traffic has been radically reduced (though is picking up more recently). We don’t know overall walking and cycling levels during lock-down, but given workplaces, schools and most shops are closed, they have likely fallen.

What happens next is massively uncertain. During the lockdown, everyone except for key workers have been staying at home. This will change as the lockdown lifts, depending on the percentage of people who continue to stay at home. There are various reasons for continuing to stay at home: working from home; looking after children at home (particularly if schools remain closed or are only partially opened); and people who are furloughed or become unemployed. Two scenarios are shown above, with 50% of people continuing to stay at home and 20% of people continuing to stay at home. The 50% figure is likely to be closer to reality for at the least the next few months, with infection rates still significant and most schools and workplaces still closed.

Using the 15% social distancing capacity limit estimated by TfL, we can see that 8 million daily public transport trips (or 4 million return trips) can no longer be made. The question then is what happens to these trips. The Active Travel scenarios shown above are based on a big increase in cycling and walking trips remaining constant or increasing. The Car Travel scenarios in contrast estimate what will happen if these trips do not switch to active travel, and significantly more driving occurs. This results in more car trips than pre-lockdown and consequent problems of increased congestion and air pollution.

How realistic is it to assume that public transport trips could switch to active travel modes? One basic way to consider this is using average trip distances by travel mode, as shown below. It is clear that National Rail trips are often long distance, and substitution with active travel is likely going to be unrealistic for most of these trips. This seems to be the assumption in the TfL plans, as the Car Free Streespace corridors prioritise linking up the biggest mainline railway stations. The idea is that mainline rail trips will change the next leg of their journey to cycling and walking, rather than tube or bus. There will likely be significantly more demand for cycle parking/hire facilities at stations and for taking bikes on trains.

The most popular public transport mode in London is the bus, and bus trip lengths are very close to average cycling trip lengths, so prospects for active travel substitution are much better for these trips. Average Underground distances of just under 10km are potentially workable as cycling trips of around 30-40 minutes (or quicker with electric modes), so some substitution is possible for these trips. The network analysis in the working paper found average trip lengths for shortest path between stations in Inner London to be just over 10km. The extent to which public transport trips can be substituted with walking trips is highly questionable for distances above 5km, as this will take most pedestrians an hour or more. The average walking trip in London is less than 1km. This likely makes cycling the key active travel mode for most public transport substitution, as well as there being significant opportunities for expanding recent micromobility options such as e-bikes and e-scooters.

Summary
The Mayor and Transport for London have proposed a radical Streetspace for London plan to redirect trips to walking and cycling as we move out of lockdown, and try to avoid a huge increase in car trips due to social-distancing capacity limits placed on public transport services. The Mayor and his team are essentially using this crisis to rapidly speed up active travel plans for London that were on the drawing board for the next decade. These plans are not empty words, they are backed up with a comprehensive transformation of many Central London routes to car free routes, as well as further restrictions on car use through the Congestion Charge. These measures are already being implemented.

At present, the proposed cycle network has lots of excellent proposals, yet remains incomplete, missing key links such as Oxford Street, as well as lacking proposals for linking Outer London centres. The proposals require support from London’s borough councils to be developed, and so far have support from key boroughs such as the City of London (with other pro-cycling boroughs such as Hackney and Camden also likely to be supportive). Hopefully more agreements will be reached with boroughs such as Wesminster, otherwise the new streetspace links will miss key parts of Inner London.

What can the next mayor do to tackle London’s hazardous air pollution?

Posted in London, SustainabilityTagged , , , , , , Leave a comment

If news emerged that a virus was killing thousands of vulnerable Londoners every year then there would justifiably be public alarm, followed by swift political action. Yet the impacts of this imaginary scenario are similar to what air pollution has been inflicting on Londoners for many years, while the political response has been slow. According to the most up-to-date and comprehensive research, air pollution reduces Londoners’ life expectancy equivalent to the death of 9,400 people of average age every year (5,900 due to NO2 and 3,500 due to PM2.5) as well as exacerbating respiratory diseases in vulnerable populations. Despite the severity of these impacts, it has become standard practice over the last decade for national government and the London Mayor to gloss-over the problem, downplay continued breaches of European law, and for politicians to delay taking any significant responses.

The current London mayor, Boris Johnson, has been slow to face up to the challenge. His tenure began by removing the western extension to the Congestion Charging Zone, thus increasing vehicle numbers in Inner West London, and opting out of using the Congestion Charge to tax the most polluting vehicles. This would have boosted the adoption of hybrid and electric vehicles (still only a meagre 1% of new cars) and reduced heavy emitters in Central London. Furthermore Johnson has not succeeded in comprehensively upgrading London’s bus and taxi fleet to hybrid and electric vehicles, instead embarking on an expensive new Routemaster project which still produces substantial emissions and currently makes up a minority of the overwhelmingly diesel bus fleet.

NO2 Annual concentration in London 2010, modelled by the London Atmospheric Emissions Inventory. All of Inner London and many major roads in Outer London greatly exceed the EU limit.
NO2 Annual concentration in London 2010, modelled by the London Atmospheric Emissions Inventory. All of Inner London and many major roads in Outer London greatly exceed the EU limit.

As recently as 2014, Johnson was still playing down the air pollution problem, in the face of strong evidence of London’s busiest streets like Oxford Street having amongst the highest concentrations of nitrogen dioxide in the world. For much of his time as mayor, he has assumed that the issue is not a public priority, and that driver and transport demands trump health concerns. This approach looks more questionable by the day. Evidence continues to mount that air quality is comparable in its health impacts to passive smoking and obesity. The consequences of the current Volkswagen scandal are still emerging as the huge gap between the advertised environmental performance of diesel vehicles and what we are really breathing in our cities is revealed. And the full costs of poor air quality are only beginning to be counted (e.g. billions of pounds added to NHS bills).

Looking to the future, London must now tackle the air pollution problem with a new mayor in the upcoming 2016 elections. The scope for change looks positive, as the main candidates on both the left and right have been openly discussing the need for change. What then could a new mayor do to address air quality and improve the health of Londoners?

Pricing Out Polluting Vehicles with the Ultra Low Emission Zone
To be fair to Boris Johnson, he did finally respond to pressure and announce a significant air pollution policy in 2013, dubbed the ‘Ultra-Low Emissions Zone‘. The ULEZ is set to come into effect in 2020 and essentially uses the Congestion Charging infrastructure to tackle the most polluting vehicles (similar to what Livingstone proposed way back in 2006). The question is whether the ULEZ goes far enough. It’s based on Euro emission standards, which measure NO2 and PM10/2.5 as well as CO2 emissions. It correctly targets diesel vehicles, requiring them to meet the Euro 6 standard, which only came into force in September 2015, so almost all current diesels on the road would have to change or pay the charge. This means that the main source of NO2 emissions will be appropriately targeted.

The current Congestion Charge boundary will be the basis of the Ultra Low Emissions Zone (map by TfL).
The current Congestion Charge boundary will be the basis of the Ultra Low Emissions Zone (map by TfL).

There are limitations however. The charge is only going to be £11.50 (except for large buses and HGVs), so it likely many diesel users will continue to drive into Central London and pay the charge. Furthermore the ULEZ will only affect Central London; the rules on diesel taxis and private hire vehicles have not been decided; and ULEZ zone residents get an overly generous exemption until 2023.

The next mayor will have a number of options that could be pursued if they want to be bolder than Johnson’s ULEZ proposal. These include increasing the ULEZ charge and/or Congestion Charge, bringing the ULEZ implementation forward to an earlier date, and potentially reintroducing the Congestion Charge Western Extension. All of these would make significant improvements to air quality, but would be controversial with drivers, many of whom were mis-sold diesel vehicles in the past as a supposedly environmentally benign option. The mayor may also want to re-examine the wider Low Emission Zone (LEZ) that tackles very polluting vehicles entering the whole of Greater London, as this affects the entire city and has not changed since 2012.

Euston Road, one of London's busiest and most polluted routes with NO2 annual concentrations typically twice the EU limit (photo by D Smith).
Euston Road, one of London’s busiest and most polluted routes with NO2 annual concentrations typically twice the EU limit (photo by D Smith).

Banning Diesel?
A more radical approach currently being proposed for Paris is to ban diesel cars all together. This would certainly make a huge improvement to air quality. The problem for London would be that the city’s bus and taxi fleet would fail this restriction, and there would likely be a backlash from thousands of car and van drivers. This would be a very confrontational approach for a new mayor.

The more pragmatic solution for London would be to work with the ULEZ framework, considering stricter measures and higher charges, and implementing the policy earlier. One very important issue for the new mayor is emerging from the current Volkswagen emissions scandal. It looks like the latest Euro 6 diesels produce significantly higher NO2 emissions in real world driving conditions compared to the misleading testing conditions. If this is the case then the proposed ULEZ will be much less effective in improving air quality (as Euro 6 diesels will be exempt from the charge). The ULEZ may have to charge all diesel vehicles, regardless of their Euro rating. This decision would need to made very early, so that there is sufficient time for car users to adjust behaviours accordingly.

Pedestrianising Streets and Upgrading the Bus Fleet
Oxford Street is London’s busiest pedestrian street and has some of the worst NO2 pollution in Europe. A response that is gaining popularity is to pedestrianise the whole street. Certainly this would be an iconic change to both the image of London, and to the experience of the city for thousands of Londoners and tourists. The challenge for this measure is avoiding creating bus jams at either end of Oxford Street, and handling the extra demand levels that would be placed on the already congested Central Line. These challenges are not necessarily permanent however, as the arrival of Crossrail in 2018 (which follows the west to east Central Line route through Inner London) has been described as a ‘game changer’ that makes pedestrianising Oxford Street a realistic option. The new mayor should certainly pursue this possibility.

The wider challenge for the mayor is not to just think about Oxford Street, but to accelerate the upgrading of the entire bus and taxi fleet towards zero emission vehicles and develop wider pedestrianisation policies. This would improve air quality for all of London’s high streets. London is still at the trial stage of pure electric double-decker buses, which represents disappointingly slow progress considering how long we have known about air pollution problems. There’s been better progress on electric/hydrogen single-deckers. Considerable investment will be needed to upgrade the bus fleet, with decision-making based on real evidence and value-for-money. Johnson’s Routemaster project has failed both these criteria, and better leadership is now required.

Heathrow Third Runway
All the main candidates for the London Mayor oppose the expansion of Heathrow, with air pollution alongside noise pollution amongst the most important factors. You can see how Heathrow already dramatically increases emissions in West London in the NO2 annual concentration map at the top of this article using the London Atmospheric Emissions Inventory data. This is due both to emissions for aircraft engines, and the thousands of cars driving to and from Heathrow everyday. The challenge will be how effectively the new mayor can challenge the UK government to reject the Heathrow expansion option.

Overall, the issue of air pollution has gained unprecedented prominence as we approach the 2016 mayoral elections. The main candidates need to respond accordingly in their manifesto commitments, with London having a lot of catching up to do after eight years of little change. It will be interesting to see which candidates are prepared to be bolder.

 

 

 

 

Understanding Household Energy Use in England & Wales

Posted in SustainabilityTagged , , , , , Leave a comment

Household energy use is a key indicator for understanding urban sustainability and fuel poverty, and is a timely topic now that winter has arrived. The LuminoCity3D site maps domestic energy use in England and Wales at 1km2 scale using data from DECC. This map has also just been published as a featured graphic in Regional Studies Regional Science. The household energy use distribution is really fascinating, with large scale regional variation and fine scale intra-urban patterns identifiable-

EnergyUse_EW_RSRS_FeaturedGraphic_web
Average domestic energy use 2012, click to view interactive map

Graph

The lowest energy use per-household is found in cities and towns in the South-West region such as Plymouth and Exeter, and also along the South coast. While the highest energy use per-household is found in commuter belt towns around London. The variation within city-regions is very high, with for example London and Manchester averages varying by up to a factor of 5, from a mere 8kWh to over 40kWh per year.

The main drivers of energy use are generally housing type (more exposed walls=more energy use; larger house=more energy use), household size, wealth and climate. Often these factors are correlated at household and neighbourhood levels- so for example wealthier households in England and Wales are more likely to live in large detached houses, and these households tend to be clustered together. These trends produce the high energy use pattern seen in London’s commuter belt, as well as in the wealthier suburbs of other large cities like Birmingham, Manchester and Leeds. South West England on the other hand benefits from the mildest climate in the UK, has a relatively high proportions of flats and generally lower average household sizes, thus resulting in the lowest energy use.

We can see how these factors play out for London in the map below. The height of the hexagons shows density, with higher density areas clearly using less energy. City centre households have considerably lower energy use, with a strong bias towards Inner East London where incomes are lower.

LondonEnergyUse
London average domestic energy use 2012, click to view interactive map

Energy use areas correlate strongly with the most prevalent housing type map (also on the LuminoCity3D site), with flats and terraced housing the lowest energy users, and detached and semi-detached areas the highest.

LondonHousingType
London most prevalent housing type 2011, click to view interactive map

The relationship with household size is less clear cut, but it can be seen that average household sizes are smaller in the city centre. On the other hand, areas with high average household sizes such as Stratford and Wembley, do not have particularly high average energy use.

LondonHouseholdSize
London average household size 2011, click to view interactive map

Overall domestic energy use patterns tend to mirror transport sustainability, in that higher density city centre areas perform more efficiently compared to low density suburbs. On the other hand the link to city size (which tends to be strong in transport sustainability relationships, with bigger cities reducing car use) is much weaker, and the most efficient locations are often small and medium sized towns and cities. It is not clear in this analysis whether more recent green policies (such as improved insulation or CHP schemes) are having much effect, but several cities with green reputations like Brighton and Bristol are amongst the best performing cities.

 

Is Developing the Greenbelt the Answer to London’s Housing Crisis?

Posted in Housing, SustainabilityTagged , , , 2 Comments

Following rapid growth and a chronic lack of new development, housing affordability has reached crisis levels in London. Median house prices are at £300k (8 times median household income) while average prices have passed half a million. London is now amongst the most expensive cities in the world, a situation with severe consequences for economic competitiveness and for inequality. Rents continue to increase faster than wages, ownership is being restricted to affluent populations and the social housing waiting list now stands at 345,000 households, nearly double the figure from 15 years ago.

Recent development figures have been very low. London needs at least 50,000 new homes per year to meet demand, yet only 21,000 were built last financial year, and this figure has been below 20,000 for all of the last five years. Nationally around 200,000 houses a year are needed, and we are building around 100,000. These figures amount to a comprehensive failure of national and mayoral policy.

Clearly substantial changes are needed. Last week the Centre for Cities outlined how this change could happen, launching their ‘Building Change: Delivering Homes Where we Need Them‘ report. It convincingly argues that we are failing to deliver homes where demand in greatest- in the vicinity of rapidly growing towns and cities- resulting in spiralling housing costs.

The report makes a range of positive recommendations for enhancing local authority capacity in relation to delivering new housing, including the streamlining and reform of compulsory purchase orders for faster development and allowing cities to benefit from uplift values in land; allowing local authorities to borrow more with longer term commitments from central government; and enabling greater cooperation between local authorities to tackle city-region challenges. Best practice examples are provided from local authorities that have successfully delivered new housing, such as Bristol and Milton Keynes. The report also provides a useful summary on brownfield capacity, with for example the potential for 350,000 homes on brownfield sites within the GLA.

Releasing Greenbelt Land for Development
By far the most politically controversial aspect of the report is the recommendation to reconsider greenbelt development restrictions. Prioritising brownfield land has been a central foundation of compact city planning over the last twenty years, directing development towards inner city regeneration and away from rural areas. Yet brownfield land can be expensive to develop, and in combination with greenbelt restrictions, land prices have soared. These spiralling land costs have significantly curtailed new housing.

Opportunities for housing on ‘Usable Greenbelt Land’ around London are mapped in the report (figure below), based on locations within 2km of rail stations. The Centre for Cities estimate that there are opportunities for 430,000 housing units on greenbelt land within the GLA, and opportunities for a massive 3 million housing units on the London greenbelt beyond the GLA boundary. This huge housing capacity could effectively solve London and the South East’s housing crisis. So is developing on the greenbelt the answer?

Opportunities for new housing on London greenbelt land, Centre for Cities Delivering Change Report 2014.

Usable Land and the Value of the Greenbelt
The gigantic housing development capacity figures quoted in the Centre for Cities report certainly demand attention. As housing development is such a central issue for planning in the South East, I have decided to repeat the Centre for Cities spatial analysis from a sustainable urbanism perspective and assess how realistic these recommendations are, and what the environmental consequences of the greenbelt development approach are likely to be.

First of all, some details on the Centre for Cities methodology. Their Usable Land definition is a 2km crow-flies buffer of rail and underground stations, excluding several environmental protection area types (SSSIs, AONB, SAC, SPA, Ancient Woodlands). The report does not argue that all this land should be developed, rather that it could be considered for development on a case by case basis. They take a ballpark figure that, given infrastructure, services and removing highly amenable land, 60% of the remaining land could be developed for housing at an overall average of 40 dwellings per hectare (thus each hectare of usable land effectively translates to 24 homes). I have repeated this method below and I get a very similar result of 120,000 hectares / 2.87 million homes on London greenbelt land beyond the GLA boundary. I get a lower (but still substantial) figure of 12,700 hectares / 306,000 homes on greenbelt land within the GLA.

SouthEast_GreenbeltDev_Map1b

There are two main spatial analysis issues with the Centre for Cities method of identifying usable land- firstly there are significant development restrictions missing, and secondly there are problems with using rail station buffers as a proxy for sustainable travel. Regarding the first problem, the most significant restrictions that should be included are flood risk areas, and additional environmental land and habitats (principally Priority Habitat Areas). The impact of these additional restrictions is shown in the map below. Surface water and flooding risk in particular covers large areas of land in the Thames Valley west of the GLA, and north in the Lee Valley, reflecting the role of the greenbelt in flood management. Assuming these areas would not be developed, this removes nearly 40% of the usable land from the analysis, leaving 75,000 hectares. With more data and time, further restrictions could be considered, for example local site access, road congestion, airport flight paths, heritage restrictions etc.

SouthEast_GreenbeltDev_Map2c

The second problem is how to consider public transport accessibility and sustainable travel. The basic principal used by Centre for Cities is sound- directing development to areas of public transport access. But locations within 2km of rail stations in the South East are often very small towns and villages, lacking local retail and services opportunities. Not surprisingly these small towns are generally highly car dependent, with around 80% of commuters driving to work, and similar patterns for other trip purposes. Building further low density housing in these locations would likely reproduce this pattern of car dependence.

Ideally the appropriate method here would be to do some accessibility modelling and network analysis (comparable to the PTAL approach used in the London Plan) to identify locations with access to local services and a range of public transport options. Unfortunately performing accessibility modelling for the whole of the South East is not trivial. The maps below shows a simpler alternative, identifying locations within an estimated local walk/bus trip of a retail and service centre (3km of a large centre, 2km of a medium centre or 1km of a small centre) based on 2010 Valuation Office data, in addition to the 2km buffer of rail stations. It is clear that a stricter definition of accessible locations greatly reduces the resulting volume of usable land, directing potential development to larger settlements with more facilities (and public transport services) like Southend, Maidstone and Hemel Hempstead. In this case it leaves 27,500 hectares of greenbelt land beyond the GLA, or 23% of the original figure. Note we also haven’t considered public transport capacity, which is a critical issue for commuters into London as many services are overcrowded.

This analysis points to the Centre for Cities figure of 3 million potential homes in the greenbelt being a big overestimate if sustainable planning guidance is going to be followed. Yet even with this stricter approach I still get a large figure of 27,500 hectares of potential development land in the greenbelt beyond the GLA, which would be about 650,000 homes at suburban densities or more at higher densities. This could go a long way to alleviating the housing crisis in the South East. The Centre for Cities report is convincing in its wider policy argument that land should be ‘evaluated on its merits’ rather than being fixed by blanket restrictions. Greenbelt development could play an important and perhaps even relatively sustainable role in addressing the housing crisis.

The question then is how any release of greenbelt land can be managed to prevent sprawl and retain the many environmental roles that the greenbelt embodies. There is also the problem of making the case to the public when the greenbelt has traditionally been a popular policy. And so we come back to the issue of local authorities cooperating to tackle regional challenges. A million commuters cross the GLA boundary every weekday, yet regional planning is almost non-existent. Any release of greenbelt land needs to be considered in its regional context and balanced against brownfield opportunities. The biggest housing opportunities are linked to new infrastructure (e.g. Crossrail both West and East of the GLA; the Varsity Line for Oxford-Milton Keynes-Cambridge), again at the regional scale. Its hard to see how the housing crisis can be tackled without much greater regional cooperation and some form of regional planning for the South East.