Panel on Urbanisation

Home > Activities > Committees > Urbanisation

Panel on Urbanisation

Rethinking the Estimation and Projection of Urban and City Populations

New York City, 9-10 January 2006. Faculty Room, Low Library, Columbia University

What Can We Learn from Remotely-Sensed Data?

Three areas would appear to be key: (1) Urban footprint detection; (2) Changes in spatial extents over time; and (3) Intra-urban classifications.

Detecting urban footprints
Can urban areas systematically be detected? What is the lower limit on the detectable city size? (By “size,” we mean both population and geographic size.) Can all types of areas (e.g., ecological or economic differences) be equally well detected?

What are the systematic biases that need to be understood before such methods are used in interdisciplinary contexts? For example, if footprints are to be based on the night-time lights, and the sensor detects permanent light (e.g., electricity or fires), what would this imply for use in less-developed countries that may have little change in urban extent but an increase in light? Can we expect there to be differences in the extent to which wealthy and poor places are detected? Or by ecosystems—e.g., coastal, mountainous, drylands, highly vegetated? For example, if a greenness-indicator (or lack thereof) were used to detect extent, such as in Landsat imagery, would urban parklands appear as urban? Or, if radar data relies on changes in angles (such as built-up areas) as the means of measurement, what biases might it have in the detection of areas that have high-slope change?

Potential to detect change over time
Over what period of time can change—whether in spatial footprints or intraurban classes—be detected? Which RS data are most useful for analyses of change? That is, which sets of data (a) could be used historically (e.g., back to the 70s); (b) will likely be collected in the near future; and (c) will likely be collected 20 years from now? Will one or many data sources be required? Is all of this detectable from a single data source? If not, can multiple sources be combined to form spatially-explicit time-series estimates of urban areas, at a global scale?

Potential for intra-urban classification
Can within-urban-area features be systematically detected? What methods are available to identify slums and other poor areas? (Likewise, suburban areas, ex-urban.) At what scale is identification possible? How might the methods be biased systematically? (E.g., can suburban areas be distinguished from park land if the suburban areas are highly vegetative?)

Measuring the Components of City Growth

Components of change
Is it advisable for estimates and projections of city population growth to make use of data on urban fertility and mortality rates? Since urban fertility and mortality survey data are rarely representative of individual cities, how should we take advantage of data that represent broader geographic areas (e.g., regions and sub-regions within countries)? What is the role of national demographic surveys such as the DHS and MICS in providing fertility and mortality data, as against censuses and other local sources of information? What would be the ideal frequency of observation? Can global estimates be assembled using a mix-and-match approach (surveys, censuses, vital registration)?

Migration
Can and should data on internal migration be included in estimates and projections of city growth? Are internal migration data available for a sufficiently broad range of countries to support the UN’s city projections? Can indirect techniques help to further expand the range of countries with usable data? As with urban fertility and mortality, data on migration are rarely collected at the city level. How should estimates based on broader geographic areas be used? Should international migration also be considered?

Urban poor
What demographic data sources can be used to identify the location and characteristics of slum-dwellers and other urban poor? How should these data be combined with data from remote-sensing techniques in identifying the spatial extent of slums?

Modeling urbanization and city growth
Is the process of urbanization sufficiently uniform across countries to allow data to be pooled for the purposes of estimation and projection? Are there important features that are unique to given countries or regions?

New Forecasting Methods

Forecasts of urbanization and city growth
Is the poor record of city and urban population projections due to a lack of data, inadequate models, or both? What distinctive contribution is made by the UN projections that adds value to, or puts in context, the projections made by local experts?

Types of Cities?
Could a taxonomy of city types help in preparing better urban projections? What would be the critical elements of this taxonomy?

Components of change
If forecasts of city population growth make use of data on urban fertility and mortality, how should we forecast these demographic inputs? Is there any basis on which internal migration can be forecast?

Forecasting spatial extents
What methods allow the spatial extents of city populations to be forecast? What has been learned from the past? Has spatial extent ever been linked to specific elements of demographic change? Should we consider investing in retrospective analyses that are generalizable rather than city-specific? What measure of uncertainty would be associated with spatial forecasts?

Scalability
Do methods for city projections scale up to the national, regional, and global levels? What sorts of checks between these levels are desirable?

Interaction
Should projections of city populations be made independently on a city-by-city basis, or be made jointly for sets of cities? What theory and methods would guide projections for sets of cities? Should national, regional, and global projections be based on aggregated city projections? If so, how do we handle towns and small cities, which generally account for an important proportion of all urban residents? How do we group cities—by population, geographic attributes, or both?

Uncertainty
Do we need new measures of uncertainty to accompany new projection methods? What levels of uncertainty would be acceptable for new approaches? (and, how do these levels compare with the uncertainties implied in older methods?)

The urban poor
What data and methods are available for forecasting the number of slum-dwellers and other urban poor? Can such forecasts be made city-specific or specific to sets of cities?

Urban Projections, Poverty, and the MDGs

Investing in data
Are the data requirements to monitor the Millennium Development Goals (MDGs) and other development goals closely related to those required to do better urban projections?

The Urban MDG
The most explicitly urban of the MDGs is Goal 7 (Ensure Environmental Sustainability), Target 11: "By 2020, to have achieved a significant improvement in the lives of at least 100 million slum dwellers." UN-Habitat is assigned a lead role in monitoring progress toward this target.

How can demographic survey and remotely-sensed data help to define, quantify, and characterize the number of slum-dwellers? Can these data be cross-validated by local observation?

Multiple urban MDGs
Other key goals and targets address health, education, water supply, and sanitation. In these areas, how important are urban and city estimates and forecasts for monitoring and reaching the MDGs? How do such urban estimates and forecasts influence the costs of achieving the MDGs?

Data availability
Which data sources are relevant to the MDGs but still under-utilized? What are the existing and emerging non-conventional data sources that should be exploited?

The changing nature of censuses
What investments are required to have spatial and tabular information from censuses remain linked in the data dissemination process? How small is small enough (in terms of administrative area size)? What does this mean for urban areas?

What can be expected from the next global census round? Will the transition to rolling census systems have a significant impact in the field?

Survey data providers
How could current surveys be made more informative about the urban landscape? Should demographic and other surveys be encouraged to over-sample urban and slum dwellers for the purpose of MDG monitoring or future population projections? Are there questions that could be added to the surveys to increase their utility when linking to or cross-validating remotely sensed data?

Remote sensing data providers
Which data streams are most likely to persist in the future? Are new sensors forthcoming that would improve our ability to detect urban features? How do we prepare basic data products from these streams that can be used by the demographic and policy communities? Is there an existing institutional home for that type of data preparation?