The Geography of Choice

2. Approaches to define rural daily systems

Abstract

The concept of “daily system” refers to a territorial system that is self-contained, where the people living in it can have access to homes, jobs, recreational spaces and services on a daily basis. These services and interests may be all located in geographic proximity, but cut across different administrative boundaries for local areas (e.g. municipalities). Although local areas may also have statistics available, they are not always the most suitable unit of statistical comparison. Functional Urban Areas (FUAs) aim to capture the spaces where people live and work, but they don’t consider people living far from cities. This chapter defines criteria for daily systems with nationwide coverage, thus covering also rural contexts. It then applies those criteria to assess existing typologies of “functional areas” (Labour Market Areas – LMAs, Bassins De Vie or “living areas” and Functional Rural Areas - FRAs). Finally, it presents and compares the main approaches that go beyond urban areas.

Ideal criteria for a typology defining (rural) daily systems

The objective is to define areas that can approximate the places where people live and work on a daily basis and that applies to all parts of a country. Table 2.1 summarises four criteria that a definition of daily systems should meet.

Table 2.1. Criteria for the definition of daily systems

Criteria	Definition
Comprehensive coverage	The delineation covers the entire population of a country.
Combines adjacent local areas	The delineation goes beyond local areas (e.g. municipal) borders to combine at least two1 adjacent areas. Every local area belongs to one and only one grouping.
Interconnectedness	A grouping contains closely interlinked places with weaker links to places outside of the grouping.
Suitable scale for daily trips	All groupings correspond to areas small enough that people can traverse them daily and large enough that people do not have to leave them for daily trips.

Comprehensive coverage ensures that the definition applies to all inhabited parts of a country, including rural areas. The definition should explicitly group all parts and not create a residual category containing unclassified parts that are potentially very different from each other.

Combining adjacent local areas is essential, as data are often available for local areas as the lowest level of aggregation. Grid-level data, available at the level of sufficiently small regular grid cells (i.e. at least 1 km resolution), are indeed becoming increasingly available, but they don’t represent the norm yet. Contiguity also matters: as the goal is to define daily systems, the local areas in a grouping should border each other. Assigning one local area to several groupings is problematic, as it would make it difficult to allocate statistics and could lead to double-counting.

Interconnectedness improves the comparability of areas within and across countries and avoids arbitrary divisions of an integrated area. Connections should be people-based and typically build on movements between places. Administrative units do not often meet this criterion. Their boundaries mostly follow historical circumstances, and highly connected local areas can end up in different groupings. For example, the work done by commuters from nearby municipalities contributes to the value of goods and services produced in Brussels. A measure of GDP per capita that divides value created within the local area borders of Brussels by the resident population of Brussels only will, therefore, be heavily inflated (OECD, 2024[1]).2

Suitable scale for daily trips should ensure that there are sensible boundaries which correspond to people’s everyday experiences. Without it, there can be groupings that are too large to realistically be crossed daily, or so small that many people will inevitably move between different groupings to work or access services. Boundaries based on interactions between places help distinguish a daily system from a group of places that simply share similar characteristics. For example, distant agricultural territories may have similar economic or physical features without being connected through everyday journeys.

Functionality is an additional criterion to evaluate typologies. Unlike the four criteria above, it is not a necessary condition for an appropriate definition of rural daily systems. However, it has an important policy dimension. Functionality describes connections concerning one or several specific functional purposes. For instance, FUAs focuses on employment as a functional criterion by grouping local areas according to work commuting flows (see Box 2.1 for more details). Sweden has a typology that takes several functional criteria into account to classify municipalities along the urban-rural spectrum: commuting patterns, but also turnover in retail, restaurants, or hotels (SKR, 2023[2]). As more data on trips become available, it is possible to build definitions that rely on the interconnectedness of local areas beyond one specific functional purpose. Mobile phone data are particularly useful, as they track trips regardless of their functional purpose.

Box 2.1. Functional urban areas

FUAs follow a functional criterion by grouping local areas according to work commuting flows. This makes the definition of FUAs a people-based approach, as it relies on people's work-related mobility patterns, rather than administrative boundaries or morphological approaches like built-up space. A FUA consists of a densely inhabited city and of a surrounding area whose labour market is highly integrated with the city (OECD, 2012[3]).

The EU-OECD framework defines FUAs using the following four-step procedure (Dijkstra, Poelman and Veneri, 2019, p. 5[4]):

1. Identify an urban centre: a set of contiguous, high density (1 500 residents per square kilometre) grid cells with a population of 50 000 in the contiguous cells.

2. Identify a city: one or more local units that have at least 50% of their residents inside an urban centre.

3. Identify a commuting zone: a set of contiguous local units that have at least 15% of their employed residents working in the city.

4. A functional urban area is the combination of the city with its commuting zone.

Note: country-specific deviations from the 15% threshold exist (e.g. Latvia applies a 20% threshold).

Evaluation of existing OECD typologies

The OECD has developed several territorial typologies, but it does not yet have a definition for all OECD countries that fully captures the concept of daily systems for entire national territories in an internationally comparable way. Table 2.2 evaluates existing typologies based on the criteria described above. They include administrative units, the degree of urbanisation and FUAs.

Table 2.2. Evaluation of existing OECD typologies

	Comprehensive coverage	Groups adjacent local entities	Interconnectedness	Suitable scale for daily trips	Functionality
Administrative units	√	√	X	X	X
Type-based definitions (e,.g the degree of urbanisation)	√	X	X	X	X
Functional urban areas	X	√	√	√	√

Administrative units such as Territorial Level 2 (TL2) or Territorial Level 3 (TL3) regions serve political and administrative purposes, but they are not ideal as statistical units for national or international comparisons of daily systems. Although these areas group smaller areas such as municipalities, they often do not capture daily systems, as their definition does not necessarily involve the criteria outlined above. Some administrative borders might coincide with the daily system of their inhabitants, but more commonly, the areas identified are too small or too big compared to daily systems. For example, people may commute beyond their TL3 region or live within an area considerably smaller than their TL2 region. Similarly, a TL3 region may contain several daily systems that span beyond its borders. International comparison is further complicated by wide differences in the size and fragmentation of TL3 regions across countries (OECD, 2023[5]). For example, the TL3 region of Paris covers only the most central part of the Paris urban area, while the TL3 region of Rome includes smaller towns and rural areas in the vicinity of the municipality of Rome (OECD, 2020[6]).

Type-based definitions, including the degree of urbanisation and the urban-rural typology developed by Nordregio (Stjernberg, Vasilevskaya and Penje, 2024[7]), classify areas, but do not delineate daily systems. While these frameworks propose typologies of local areas, units in the same grouping are not required to be adjacent, and groupings will often be scattered across entire countries. The degree of urbanisation, for example, can be used to identify rural municipalities within a country, while considering each municipality independently from each other. While type-based definitions provide useful information for certain types of targeted policies with a centralised, nationwide scope, they do not allow for differentiation between rural areas in different parts of a country (e.g. near a large city or in a remote area), limiting their relevance for locally targeted policies.

Combining administrative units and type-based definitions yields additional insights, yet it does not eliminate the shortcomings of the two inputs. For example, Fadic et al. (2019[8]) classify OECD TL3 regions regarding their access to large FUAs. Doing this allows the computation of statistics by the type of place, while keeping a local dimension for targeted policies. Nevertheless, the classification still considers individual TL3 regions and thus the issues related to administrative definitions whose boundaries may not match daily systems, remain. A comprehensive definition of daily systems can improve this. Once such a definition is in place, type-based definitions can provide further detail into the characteristics of the areas. For a recent application to Labour Market Areas in Italy, see Franconi, Mantuano and Ichim (2024[9]).

Finally, FUAs fulfil the criteria of suitable scale for daily trips, interconnectedness and functionality. A FUA describes a daily system that consists of a dense urban centre defined from a population grid with at least 50 000 inhabitants and surrounding local areas connected to that centre via work commuting (Box 2.1). FUAs typically contain multiple local areas but can cross the boundaries of higher-order administrative units. They are already well established and considerably improve the national and international comparability of urban areas. However, given their definition around city centres, FUAs lack comprehensive coverage, as they do not take most rural areas into account by design. Daily systems that cover the entire territory of a country are thus still missing.

Given that FUAs are already defined, an important question that arises is whether daily system definitions should take FUAs as given and delineate the remining space. On the one hand, a unified definition of daily systems that treats all parts of a country the same is appealing for simplicity and consistency. Furthermore, administrative-based FUAs include rural parts that may have functional connections to their surrounding areas. A daily system definition that takes FUA as given and delineates only non-FUA territories does not consider these possible connections. On the other hand, keeping the well-established definition of FUAs untouched and implementing an additional delineation strategy only for areas not classified as FUAs could be beneficial, particularly regarding the production and comparability of statistics over time. The definitions presented below are implementable for an entire country (or even across country borders) or for all country parts not classified as a FUA.

Comprehensive definitions of daily systems

There are mainly three approaches to define functional areas for entire territories: Labour Market Areas (LMAs), Bassins De Vie and Functional Rural Areas (FRAs) in combination with FUAs.

All three approaches group local areas into daily systems based on trips, but the rationale for these trips differs amongst them.

• Labour Market Areas (LMAs) focus on actual trips related to work commuting, inferred from people living and working in different local areas.

• Bassins De Vie (or “living areas”) use theoretical trips between people’s homes and existing service locations, simulated on the road network.

• Functional Rural Areas (FRAs) consider theoretical trips between people’s homes and population centres, simulated on the road network.

This report refers to LMAs as a class of similar algorithms that all share the same underlying concepts.3 Since Coombes, Green and Openshaw (1986[10]), many researchers and organisations applied a version of this algorithm to delineate daily systems.4 Bassins De Vie provides a good example for a service-based methodology, as it offers a transparent definition to partition a country into contiguous and non-overlapping areas based on access to services. FRAs is a new methodology that builds on population grids and driving time simulations on road networks, data that only became available more recently on a global scale.

Table 2.3 presents an overview of the similarities and differences between the three approaches. All three approaches fulfil the four criteria outlined in Table 2.1. The remainder of this chapter will focus on this comparison, while Chapter 3 will present the methodology for each in more detail.

Table 2.3. Comparison between Labour Market Areas, Bassins De Vie and Functional Rural Areas

Criterium	Labour Market Areas	Bassins De Vie	Functional Rural Areas
Currently delineated by	Australia, Canada, Denmark, France, Germany, Hungary, Italy, Japan, Mexico, Netherlands, Norway, Sweden, Switzerland, United Kingdom and the United States from official sources	France, similar definitions exist in other countries	New proposal by the Joint Research Centre of the European Commission, covering Europe
Requires data on	Commuting flows between local areas Size of working population by local area Administrative boundaries of local areas	Location of services and their type Road network to compute travel times Population grid Administrative boundaries of local areas	Road network to compute travel times Population grid Towns and villages from DEGURBA
Source of trips	Actual trips, Inferred from census questions or administrative data	Theoretical trips, simulated based on road network	Theoretical trips, simulated based on road network
Nature of trips	Only work-related	Only trips to service locations	Trips to population centres, not limited to one functional purpose
Origin of trips	Homes (local areas)	Homes (grid cells)	Homes (grid cells)
Destination of trips	Workplaces (local areas)	Services (geographic coordinates)	Population centres (groupings of grid cells according to a set of criteria)
Mode of trips	Many different modes possible	Currently only car travel	Currently only car travel
Maximum trip length	No hard cap	No hard cap	Combination of areas around rural centres currently capped at 60 minutes, no hard cap concerning trips to the next rural centre
Building blocks for algorithm	Local areas	Local areas	Population grid pixels
All building blocks are treated the same	Yes	Yes	No, special role for towns and villages defined by DEGURBA5
Can aggregate statistics at the level of	Local areas	Local areas	Local areas (FRAU) Grid cells (FRAG)
People-based	Yes, but only working population	Yes	Yes
Geography-based (beyond whether areas border)	No, agnostic about the means and duration of observed commuting	Indirectly via trip duration, special rules preventing to merge adjacent areas separated by mountains	Indirectly via trip duration, special rules for inhabited pixels without access to marked roads
Covers entire territory	Yes, combination with FUAs possible	Yes, combination with FUAs possible	In current version only in combination with FUAs
Restricted by regional borders	No	No	No
International version possible	Yes, if consistent cross-border commuting data are available and common parameter values are set	Currently difficult, would require internationally standardised data on service types and locations	Yes, if common parameter values are set
Parameters	Minimum size of working population Target size of working population Minimum level of self-containment Target level of self-containment	List of relevant service types Minimum number of service types for service hubs	Maximum travel time to identify rural centres Maximum travel time to combine big catchment areas Maximum travel time to combine any catchment areas Population threshold for big catchment areas
Parameters vary by	Country, uniform within countries (except France)	One set of parameters for France	One set of parameters for Europe
Manual adjustments	Necessary but limited	Necessary but limited	Necessary but limited

Interconnectedness in an international context

Each of the three definitions uses a version of interconnectedness to define daily systems that span a country's entire territory. They can aggregate statistics at the level of municipalities while being independent from higher-level administrative units i.e. the resulting boundaries do not need to follow those of larger administrative units, such as provinces or regions, and a daily system may therefore extend across them. All three definitions build on trips, but the rationale behind these trips varies substantially.

The LMA algorithm uses bilateral commuting flows6 between municipalities to delineate areas such that a high proportion of people who work in an area also live there, and a high proportion of (working) people who live there also work there. Unlike the algorithm that defines FUAs, it does not require a core-periphery setup.7 Instead, it treats all local areas uniformly ex-ante and considers commuting flows between all pairs of local areas.

However, work commutes may account for a minority of trips in daily life, especially in rural areas. For instance, Couture, Duranton and Turner (2018[11]) report that only about a quarter of trips in the largest 100 metropolitan statistical areas of the United States are work-related. Other important reasons for leaving home to go somewhere include shopping, school, medical appointments, or meeting with friends and relatives. The weight of non-work-related trips may be higher in rural areas compared to cities. In addition, employment may be more dispersed in rural areas compared to urban areas, and more people may travel to towns to consume services than to work there (Dijkstra and Jacobs-Crisioni, 2023[12]).

An ideal definition of daily systems should work for different generations, and focusing entirely on work-related trips may lead to the exclusion of parts of the population. People of every age group require access to at least some of the services and the geography of service locations can shape mobility patterns in rural areas. For example, education-related trips are likely to be relevant for higher education levels such as upper secondary schools, which are often concentrated in larger rural settlements (OECD, 2021[13]). Retired workers and old-age people still often travel to in-person services like banks, pharmacies, or supermarkets.

The Bassins De Vie (Insee, 2024[14]) approach focuses on service locations, classifying local areas as service hubs if they host many different types of services. It then assesses where people in the remaining local areas will likely obtain these services by simulating driving times from their homes to the locations of the services that are missing in their local area. A local area and a service hub are grouped together if that service hub provides the people in the local area with the fastest access to more missing services than any other service hub.

The Bassins De Vie method assumes that people choose the closest service location, instead of observing actual trips. This assumption can be particularly problematic for expensive private sector service providers, for which affordability rather than driving time can limit accessibility. Moreover, while people often commute to their jobs five days per week, some services only require infrequent trips. This implies that an approach that tries to capture mobility for different age-groups should ideally take into account not only differences in the types of trips but also in the frequency of trips.

The FRA method, in turn, simulates generic trips from homes to population centres such as villages and towns. It then assigns grid cells to population centres, taking driving time and population size into account. Dijkstra and Jacobs-Crisioni (2023[12]) note that population centres often host services, but the presence of services is not part of the criteria for classification (services may be instead considered for analysis once the Functional Rural Areas have been delineated). The FRA definition is thus compatible with other channels of interconnectedness beyond services. In this sense, the FRA definition does not depart from a specific or explicit functional criteria, and it relies instead on the assumption that different functions happen in connection to trips linked to population centres.

However, locations that do not host the largest resident population in their surroundings can still function as the centre of daily systems. Regulations, geographic constraints and land prices might all limit the number of residents in a place that many people commute to daily. A prominent example of places that offer more jobs and services than homes for the resident population are towns or villages hosting tourist attractions. In contrast, not every cluster of large residential buildings functions as a centre.

Data requirements for international comparability

The three methods described in this chapter rely on different sources of information and thus bring different challenges for international comparability purposes.

For LMAs, the main data requirement for cross-country applications are reliable data on commuting flows. As the next chapter will argue, there are substantial challenges with this kind of data even if the concept of commuting flows generalises relatively well across countries. Moreover, the parameter value selection complicates international comparability. The algorithm relies on four parameters: minimum and target values for workforce size and self-containment, with the latter referring to the proportion of people who work and live in an area. There is no standardised procedure on how to set the parameter values and the thresholds chosen for different countries vary substantially. Optimal parameter values may vary systematically with some variable, for instance population density. This could be potentially incorporated into the algorithm or serve as a basis to coordinate internationally parameter value selection.

The challenges with an internationally standardised definition of Bassins De Vie are somewhat different, but probably harder to overcome. Extending the definition to work across countries would require internationally comparable data on the location of services, including services offered by the state and private companies. However, even within countries, reliable and complete data on services and other points of interest are hard to obtain. Moreover, while some service locations are comparable across OECD countries in their functioning and frequency, others have characteristics that are country- and context-specific. For example, a multitude of small shops and practices might prevail in one context, while people may get the same services in big supermarkets and clinics in another environment. If satisfactory data are available, an international classification would require agreement on the types of services included and the threshold number of service types a local area must host to be a service hub.

The FRA method already relies largely on internationally comparable sources, but it is unclear whether one set of parameters is reasonable across all OECD countries. Population grids and road data are available across all OECD countries and beyond. Researchers can choose between different data products, but they face the trade-off of choosing between the most accurate option for each country and one of the products with global coverage that are methodologically more consistent but cannot offer the same level of customised optimisation to each individual country. While the building blocks of FRAs are initially grid cells, it is possible to define additional rules to group areas defined by administrative or other criteria (e.g. municipalities). As will be discussed in the next chapter in more detail, constructing FRAs requires three different parameters for travel time and a population size parameter. FRAs have not yet been calculated by national statistical offices, however, Dijkstra and Jacobs-Crisioni (2023[12]) propose a unique set of parameters for the whole European Union. Whether these parameters work for all OECD countries, including countries with vast and sparsely populated territories (such as Australia, Canada and the United States), with very high population densities (such as Korea), or with varied access to government services, needs testing.

Finally, all three methods require some manual adjustments. They follow transparent and replicable rules, but some limited adjustments are necessary, mainly in relation to the wish to have contiguous areas. Additionally, the algorithms might not be able to assign all inhabited areas (LMAs) and require manual processing of a reserve list, or an individual grouping might become unrealistically big (Bassins De Vie) or small (FRAs).

Daily system definitions using mobile phone data

Mobile phone data offer an exciting perspective for internationally consistent definitions of daily systems (Dong et al., 2024[15])8 They have the potential to facilitate the measurement of interconnectedness between local areas and make it more precise while allowing the tracking of trips beyond one functional purpose. Mobile phone data can also inform about travel patterns of different socio-economic groups (e.g. youth, old-age people) and help test whether geographical proximity indeed determines service location choices (Nilforoshan et al., 2023[16]). With sufficiently disaggregated data, it is possible to distinguish repeated everyday trips from sporadic travels for different socio-economic groups. These trips can then serve as a basis for combining local areas into daily systems. The principle remains that most trips should happen within a given daily system, rather than across them.

An increasing number of studies use mobile phone data to study mobility patterns. Existing studies covering a total of 16 European countries used mobile phone data to delineate areas applying network analysis and showed that the number and size of the identified areas varied in different stages of the COVID-19 pandemic (Iacus et al., 2020[17]) (Osorio Arjona, Ruíz Santacruz and de las Obras-Loscertales Sampériz, 2023[18]).9 Further analysis is needed to assess the extent to which such approaches can be generalised, also considering data accessibility and privacy concerns.

Complementing the three approaches discussed in this report with mobile phone data can be a way forward worth exploring. Instead of commuting flow data from traditional sources, the LMA method can also take trips measured from mobile phone data as inputs. The Bassins De Vie approach could profit from data indicating where people actually get their services. Such data can come from mobile phones or credit cards. Finally, the FRA approach could define rural centres as places where many people spend their daytime instead of where they reside. In this case, mobile phone data could identify such centres without tracking individual people across time and space.

Comprehensive coverage for mobile phone data is far from a reality for research today, even if mobile phones are almost ubiquitous in OECD countries for all but the oldest cohorts. Currently, there is a large gap between the mobility flows data that exist and the ones that are accessible for analysis. Mobile phone companies can be reluctant to share them, or not allowed to do so because of regulations. Legal privacy concerns will continue to make research based on individual data very complicated, even when anonymised. Aggregated data on the number of people frequenting a location or travelling across places can be easier to access and already provide beneficial insights. The increased number of studies suggests that mobile phone data might soon offer an alternative for the delineation of daily systems beyond work commuting in individual countries or regions. However, reaching a point where data availability is good enough for all (or most) OECD countries, such that it can be the basis for internationally standardised and comparable daily systems, will require more access, time and much effort.

Meanwhile, mobile phone data for an individual country or region can serve as a reliable basis for validation exercises for international definitions. Daily systems delineated with other methodologies should remain stable when using trips or centres identified using mobile phones rather than administrative data. Moreover, they can facilitate comparisons between different methods by providing insights into the number of actual trips within and across areas delineated with different approaches.

References

[10] Coombes, M., A. Green and S. Openshaw (1986), “An Efficient Algorithm to Generate Official Statistical Reporting Areas: The Case of the 1984 Travel-to-Work Areas Revision in Britain”, The Journal of the Operational Research Society, Vol. 37/10, p. 943, https://doi.org/10.2307/2582282.

[11] Couture, V., G. Duranton and M. Turner (2018), “Speed”, The Review of Economics and Statistics, Vol. 100/4, pp. 725-739, https://doi.org/10.1162/rest_a_00744.

[12] Dijkstra, L. and C. Jacobs-Crisioni (2023), Developing a definition of Functional Rural Areas in the EU, https://publications.jrc.ec.europa.eu/repository/handle/JRC135599.

[4] Dijkstra, L., H. Poelman and P. Veneri (2019), “The EU-OECD definition of a functional urban area”, OECD Regional Development Working Papers, No. 2019/11, OECD Publishing, Paris, https://doi.org/10.1787/d58cb34d-en.

[15] Dong, L. et al. (2024), “Defining a city — delineating urban areas using cell-phone data”, Nature Cities, Vol. 1/2, pp. 117-125, https://doi.org/10.1038/s44284-023-00019-z.

[8] Fadic, M. et al. (2019), “Classifying small (TL3) regions based on metropolitan population, low density and remoteness”, OECD Regional Development Working Papers, No. 2019/06, OECD Publishing, Paris, https://doi.org/10.1787/b902cc00-en.

[9] Franconi, L., M. Mantuano and D. Ichim (2024), “Population grid and location quotient of land cover to capture the urban-rural nature of labour market areas in Italy”, GeoJournal, Vol. 89/1, https://doi.org/10.1007/s10708-024-11000-1.

[17] Iacus, S. et al. (2020), Mapping Mobility Functional Areas (MFA) using Mobile Positioning Data to Inform COVID-19 Policies, https://doi.org/10.2760/076318.

[14] Insee (2024), Base des bassins de vie 2022, https://www.insee.fr/fr/information/6676988.

[16] Nilforoshan, H. et al. (2023), “Human mobility networks reveal increased segregation in large cities”, Nature, Vol. 624/7992, pp. 586-592, https://doi.org/10.1038/s41586-023-06757-3.

[1] OECD (2024), OECD Territorial Reviews: Brussels-Capital Region, Belgium, OECD Territorial Reviews, OECD Publishing, Paris, https://doi.org/10.1787/0552847b-en.

[5] OECD (2023), OECD Regional Outlook 2023: The Longstanding Geography of Inequalities, OECD Publishing, Paris, https://doi.org/10.1787/92cd40a0-en.

[13] OECD (2021), Delivering Quality Education and Health Care to All: Preparing Regions for Demographic Change, OECD Rural Studies, OECD Publishing, Paris, https://doi.org/10.1787/83025c02-en.

[6] OECD (2020), Delineating Functional Areas in All Territories, OECD Territorial Reviews, OECD Publishing, Paris, https://doi.org/10.1787/07970966-en.

[3] OECD (2012), Redefining “Urban”: A New Way to Measure Metropolitan Areas, OECD Publishing, Paris, https://doi.org/10.1787/9789264174108-en.

[18] Osorio Arjona, J., J. Ruíz Santacruz and J. de las Obras-Loscertales Sampériz (2023), “Mapping of functional areas in Spain based on mobile phone data during different phases of the COVID-19 pandemic”, Journal of Maps, Vol. 19/1, https://doi.org/10.1080/17445647.2023.2214804.

[2] SKR (2023), Classification of Swedish municipalities, https://skr.se/en/skr/tjanster/kommunerochregioner/faktakommunerochregioner/kommungruppsindelning.2051.html.

[7] Stjernberg, M., A. Vasilevskaya and O. Penje (2024), Towards a grid-based Nordic territorial typology, https://doi.org/10.6027/R2024:91403-2503.

Notes

← 1.  In some cases, large individual local areas can constitute a grouping by themselves.

← 2.  The issue gets aggravated by the fact that Brussels is also its own Territorial Level 2 (TL2) region.

← 3.  Some literature uses the term “travel to work areas” to refer to the same kind of algorithms.

← 4.  The methodological description in this paper follows Istat (Franconi, Mantuano and Ichim, 2024[9]) and OECD (2020[6]).

← 5.  The FRA method takes towns and villages exogenously from the degree of urbanisation and therefore treats grid cells within these settlements different than other grid cells. However, DEGURBA classifies towns and villages from the same population grids that underpin FRAs. An algorithm that contains both the DEGURBA and the FRA classifications would therefore treat all building blocks the same. Of course, such an algorithm would be unpractically lengthy in practice.

← 6.  The recent increase in teleworking possibilities weakens the imperative of proximity between home and workplace, but this only impacts a part of the workforce and physical presence is often required for some days of the week nevertheless. Moreover, teleworking possibilities might be less prevalent in rural areas than in cities.

← 7.  FUAs also describe commuting zones / labour market areas. However, FUAs are core-based functional areas, while the LMA algorithm presented in this paper creates non-core based functional areas.

← 8.  Dong et al. (2024[15]) focus on the delineation of cities or urban areas, but their argumentation also applies to daily systems more generally.

← 9.  Osorio Arjona, Ruíz Santacruz and de las Obras-Loscertales Sampériz (2023[18]) identify between 23 and 36 areas, while Iacus et al. (2020[17]) report 188 areas before and 330 areas after the lockdown. This discrepancy shows that the choice of the algorithm and parameters is not insignificant, even when the principle, context, and data are all similar.