Nordia Geographical Publications 42: 2, 39–52 Mikko Tervo et al.
Accessibility analysis of public services in rural areas under restructuring
Mikko Tervo1,2, Ossi Kotavaara1, Harri Antikainen1 and Jarmo Rusanen1 1Department of Geography, University of Oulu 2City of Kokkola
Abstract: Finland is currently undergoing a restructuring of the municipal and public service structure. This restructuring is an effort to react, in addition to tightening economy, to the changing circumstances with regard to the demographic structure, especially in rural areas. This paper examines the use of geographical information systems (GIS) to analyze the spatial accessibility of public services and unearth differences in the accessibility of different demographic subgroups in a rural municipality with a developed road transport network, but declining population and economy. The potential for applying GIS methods to future rural development and policies, in light of an increasing trend of municipal merges, is also discussed. This paper seeks to show that GIS driven demographic analysis are valuable in exposing equality issues and differences between possible locations when provision of public services is planned. Datasets used in this research are digital road network data covering the full extent of the Finnish road network and population grid data including the population of Finland in various demographic subgroups on a 1 km × 1 km grid cell size. Such datasets offer a unique insight into the spatial division of population and transport possibilities. Accessibility varies for different demographic groups and this can be projected and numerated by the use of two GIS methods: relative accessibility measures and average distance measures.
Introduction population (Tervo 2005). Thus, Finland’s population centers are generally separated Finland, with the majority of Europe, by long distances. Finland is very peripheral is in the final stages of demographic in the European context not only by development, also facing a state of geographic location, but also by different degeneration in peripheral populations with peripherality indexes (Schürmann & Talaat a strong trend of urbanization highlighting 2002). Also, it is noteworthy that areas the struggle of maintaining affordable that are peripheral on the Finnish national public services in rural and peripheral areas. scale can be seen as extremely peripheral Finland however has some characteristics on the European scale (Spiekermann & that differ from the central areas of Europe. Aalbu 2004). Generally this means, that It is located in the north-eastern fringe while discussing European rural areas as of the European Union and is a large a whole, Finnish rural areas are arguably country in relation to its small, scattered different in terms of population density
39 Accessibility analysis of public services in rural... NGP Yearbook 2013 and peripherality. This is to say, that while policy makers, municipal decision makers there are peripheral areas around Europe, and planners to make subtle shifts into the those areas can be seen as more central if spatial formation of public services. looked upon from the viewpoint of Finnish rural areas and periphery. Also it is to be Periheral services and noted that population and services are often regional structure clustered, and peripheral areas often have low population and relatively poor access Generally, municipalities in Northern, to services (Cromley & McLafferty 2002). Eastern and Central Finland are large in This results in challenges in maintaining an terms of land area, which can be also affordable public service structure in rural seen in the map (Figure 1). In addition areas. Of course, the notion of access to to this there are many municipalities with services and every day possibilities being very small populations (Statistics Finland a challenge in rural areas is not a new one 2010). The municipal structure of Finland (see for example Moseley 1979). While the is however changing. The population is field has been visited before, the changes moving into centers while the rural areas happening in the municipal structure are declining and facing a set of future and recent developments in geographical challenges. Rural areas are declining in information systems (GIS) techniques terms of total population and also the and data available offer an interesting field demographic structure is transforming in to study. This paper also sets to utilize such a way that the number of the elderly the chance of using spatially accurate is increasing while the number of children demographic data as the population data is declining. Demand for basic services input and more importantly the chance of is changing in different sectors, such as analyzing spatial accessibility of services healthcare and education. The change is for different parts of the demographic: the such that the need for some services is youth, the working adults and the elderly. growing while others are declining. This As Tanser et al. (2010) put it: “Optimal is especially exemplified in areas with low locations can be modeled to improve existing number of children and high numbers systems, while this is potentially useful, it might of elderly citizens. Thus while the need not lead to implementation because of the huge for childcare might be declining, the need costs involved in changing expensive and entrenched for care for the elderly might be crowing health care systems.” While the quote discusses drastically. While Finland as nation is health care, the issue is the same with becoming increasingly urbanized, in the other forms of services. This is why the context of the European Union it has primary aim of this paper is to analyze many areas that can be characterized as the accessibility of different public service very rural and of a low population density. locations already in place. This means that This means travel distances are most often rather than trying to allocate and completely long in the European context and there are rearrange regional systems, this paper vast uninhabited areas between population seeks to produce information for regional clusters. There are some 78,000 kilometers
40 Nordia Geographical Publications 42: 2, 39–52 Mikko Tervo et al.
Figure 1. Finnish municipal system and the study area of Siikalatva. of public roads in Finland and most of Finnish governments have maintained a the private transportation in Finland takes positive attitude towards restructuring of place via private car, even more so in rural local governance and services since 2005 areas where public transport possibilities (see Virkkunen 2013). This has resulted in are limited. various municipal merges in yearly basis, In Finland, local municipal authorities which is arguably a very special set of events have the responsibility to provide basic involving a large number of citizens. The services to citizens. Two consecutive number of municipalities has gone down
41 Accessibility analysis of public services in rural... NGP Yearbook 2013 from 432 to 320 in the span of eight years. spatial distribution characteristics of the In the near future, even more municipal population in the Siikalatva area. merges are very likely. Merges can lead to situations in which public services must be reorganized by cutting overlapping services Reseach design and centralizing them, as economic issues are often a driving force behind the merging The aim of this paper is to examine GIS process. For the municipalities struggling as a tool for analyzing the accessibility of with the economic issues following the public services in a rural municipality that decline of rural areas ensuring proper has gone through a municipal merge and public services for its citizens can be a now consists of more than one roughly challenge. On this background, information same-sized center. Access to public services on how accessible public services are for is calculated for the municipal centers different parts of the demographic is very for the youth, the working aged and the valuable for planners and decision makers. elderly. GIS have been traditionally used in In this paper, the municipality of assistance of decision making in location Siikalatva is used as a case of a rural allocation problems in the field of business municipality with a recent municipal facility planning (for example Clarke & merge, which took place in 2009. Siikalatva Rowley 1995; Cheng et al. 2007; Yu et al. is a municipality having a polycentric 2007; Church & Murray 2009) and there structure and Spiekermann and Aalbu are some examples in which GIS have been (2004) consider it to be extremely peripheral used for service planning (for example in terms of accessibility in the European Tanser et al. 2010). context. Siikalatva is a municipality of There are several advantages in using GIS approximately 6300 inhabitants living in an as a tool to analyze public service related area of 2230 square kilometers. This equates issues, as listed by Richards et al. (1999), GIS to a population density of 2.8 inhabitants data from various sources can be combined per square kilometer. However, as is the and linked; GIS provides several new types case in most of Finland, the population of data; new GIS methods can be added is clustered in municipal centers and the as tools to be used by the planners; maps municipality has a lot of unhabited land produced using GIS can illustrate issues area. Siikalatva has four municipal centers: more efficiently than graphs and tables. Rantsila, Pulkkila, Kestilä and Piippola, that While Richards et al. (1999) discuss GIS each used to be separate municipalities but with a focus on healthcare; the same GIS are now merged and form Siikalatva. The approaches can be used for different forms map of Siikalatva including the population of services. data gives a good overview of the study It is understood that applied accessibility area (Figure 2). The classification of research can be used in planning when populated areas is based on Rusanen et provisioning social services (Kwan et al. (2004), with adjustments to the class al. 2003). Quite often GIS methods names by the authors to better display the for analyzing accessibility are used in
42 Nordia Geographical Publications 42: 2, 39–52 Mikko Tervo et al.
Figure 2. Demographic characteristics and the major roads of Siikalatva area. urban planning (for example Liu & Zhu The aim of this paper is to touch on 2004) and often in rural areas without a the set of intertwined issues of policies, developed transportation network, such political issues, demographic trends, as in developing countries. In this paper, regional development and change in however, while the area of focus is by all municipal structure and public service (re-) means rural, the road transport network is organization with GIS and demographic quite well developed and fulfills the role of analysis by seeking to answer two questions. a major means of transportation. The first question this paper seeks to answer
43 Accessibility analysis of public services in rural... NGP Yearbook 2013 is, do the four municipal centers in Siikalatva the spatial locations of each of the segments differ in terms of network accessibility starting and ending points. Thus the dataset considering different age groups and if so, can be used to estimate travel times with the how? From there on, the second question assumption that one travels at the highest is, are relative accessibility measures derived legally possible speed when aspiring to from potential accessibility measures and reach public services. The road network average distance measures useable in a rural used in the analysis involves Siikalatva area setting in order to analyze accessibility to with a buffer of 20 kilometers outside the public services? municipal boundaries to involve paths that might take place outside the municipal borders and return inside the municipal Data borders before reaching the destination.
The population database used in this paper consists of 1 × 1 km grid cells and includes Theory and calculation a wide array of social and socioeconomic variables and most importantly considering Accessibility in a rural setting this paper; the population is also classified into demographic groups. (Statistics Finland For access, there are arguably a plethora of 2010) The data is constructed from various indicators and definitions. Ricketts (2010) civil registers and is updated yearly. The has listed some concerning healthcare, population of each cell is derived from the definingaccessibility is somewhat of an easier actual number of residents in the addresses task. Accessibility is simply put, the ease by inside the cell area. This means the data which a place can be reached from another is very accurate when it comes to spatial place (Liu & Zhu 2003). To be more precise, structure of the demographic and more accessibility is the measure of the capacity reliable than datasets constructed from of a location to be reached by, or to reach censuses or estimates. The data variables different locations (Rodriguez et al. 2009). used in this paper are total population, In this research accessibility is seen as the youth (population under 20 years of age), extent to which the land-use system enables working aged (population from 20 to 63 of individuals to reach their destinations, age) and the elderly (population older than which are in this case public services (see 63 years of age). Geurs & Ritsema van Eck 2001). In this The digital road network data (DigiRoad) case the form of access measured is in its used in this research portrays the full essence spatial access, meaning that the focus extend of the Finnish road network in is on spatial distance variables (length or metric accuracy. The data is maintained by time), rather than for example social issues the Finnish Road Administration (2010). (see Lou & Wang 2003 for more), which The dataset includes speed limits for each could (and by all means should) be a field section of the road network and the length studied elsewhere. However, we argue that of the sections, which can be derived from GIS works very well as a methodology
44 Nordia Geographical Publications 42: 2, 39–52 Mikko Tervo et al. for evaluating spatial access using a more the basis of the network topology. The numeric approach. factual center (or the ‘main street’) was As a basis and an enabling factor on identified as the assumed location of public which the analyses are made, an assumption services. While identifying the actual center concerning public service is made. It is or assumed public service locations might assumed that in a rural setting the vast be difficult or impossible in an urban majority of services are located in municipal setting, it can be done quite reliably in centers. Therefore, the destinations used very sparsely populated rural areas, where in the GIS analyses are located in the the municipal centers stand out from the municipal centers, which were identified by spatial demographic surroundings and the the researcher by viewing the population road network thickens in a very noticeable grid data and the road network data. The way, as seen on the map displaying the road destinations were selected from the grid network topology of the Siikalatva area cell of the highest population and from (Figure 3).
Figure 3. Excerpt of the road network shows that the centers are easily indentified.
45 Accessibility analysis of public services in rural... NGP Yearbook 2013
Relative accessibility measures of the potential accessibility analysis is the ability to differentiate between Relative accessibility (also referred as the clustered population concentrations potential accessibility or market potential) and isolated concentrations, as well as is arguably one of the most important peripheries (Geertman & Ritsema Van practices of analyzing accessibility and has Eck 1995; Rodriguez 2006; Spiekermann its roots in the history of spatial science (see & Wegener 2007). Potential accessibility for example Harris 1954). The accessibility of the population for each location can be analytical perspective of this study is, how calculated by dividing the population count accessible is the select destination from all of of another location by the distance attribute the possible origins? rather than asking, how such as travel time between the locations, and accessible are all the possible destinations of a summarizing these values. The calculated system in relation to each other? value indicates how each location can reach The accessibility values are standardized an attracting attribute in other locations into such a form, that the most accessible related to a transport friction. Hence, the location gains the value one and the less clustered population concentrations as well accessible destinations gain values smaller as isolated concentrations and peripheries than one and are theoretically asymptotical can be differentiated. The equation for this to zero. Generally, different destinations in the simplest form can be presented as: can be easily and unambiguously compared, because relative accessibility values do not tend to fall very close to zero in practice. It is to be noted however, that in this approach, accessibility values can only be compared where A(P) is the potential accessibility for each of the destinations between the matrix, dij is the distance between the set demographic group, meaning that the location i and j, pj is the attribute of the values that can be compared against are on related destination location, n is the number the horizontal rows of the table, in which of origins, and α is the parameter for the the results are presented. transport friction indicating the efficiency As relative accessibility has its core in of the transport system and the interest potential accessibility, understanding how to move. An increase of α will lead to a potential accessibility analysis works is greater distinction between nearby and crucial in order to fully understand the distant destinations. Ideally, α should be relative accessibility values produced in estimated empirically, but usually, like in the analysis. Potential accessibility is a the case of Siikalatva, there is no adequate measure to describe how the population survey information available to do this. The of other locations can be accessed from value is highly dependent on scale and the each location involved in the analysis by type of activity modeled. At a local level, the transport network. By the analysis the higher values of α are typically used, but in centrality and peripherality of locations national and international scale analyses α is can be related numerically. The benefit assumed to be 1 (see for example Gutiérrez
46 Nordia Geographical Publications 42: 2, 39–52 Mikko Tervo et al.
2001; Holl 2007). In this case, values 1 times distance s from origin j to location i and 2 were used for α which were applied divided by the total population P (or in the also by Kotavaara et al. (2012). The former case of different demographic groups, the was used to denote linear friction increase total of the population of the demographic involving “wider” potential accessibility, groups in question). Empirically calculated while the latter was used to estimate more network distance values are applied to the locally focused potential accessibility with distance value to control the movement quadratic friction increase. connecting the grid cells into the actual The potential accessibility analysis is road network. As with relative accessibility, often considered as one of the most average distance values are also presented essential accessibility analyses in theoretical as actual length of travel in kilometers and terms, but it is used quite rarely in practice, also in standardized form. Average distance which arguably is due to the requirements values are also standardized as such that the related to the GIS techniques. To the best location of the longest average distance of our knowledge there is no commercial gains the value 1. software available for carrying out potential accessibility analyses. This being the case, in this study a Python script was applied Results and discussion in ArcGIS in order to compute needed Relative accessibility and average calculations. distance values
These analyses show that Pulkkila is the Average distance measures most accessible municipal center in the Siikalatva area (Table 1). There are however The second measure used is average some variations and interesting differences distance. The idea is to measure the average in the accessibilities of the other three distance one person has to travel via road centers, especially when different age network to reach services. In this calculation groups are involved in the analysis. This the population weight in different habited means putting the municipal centers into grid cells is taken into account. Thus, the an order from the most accessible to the calculation is not merely of the average least accessible is a task in which careful distance to habited areas, but for a single consideration is needed. person. The average distance a citizen has to For more detailed consideration, the travel to a select location can be calculated relative accessibility values in table 1 with the following formula: are shown following a pattern of each demographic group being listed first with α having the value of 1 and then with α having value of 2. Standardized values are listed after the non standardized ones. As
Where the average distance si to location can be seen from the table, when α has a i is the sum of all destination’s population p value of 2, the differences in accessibility
47 Accessibility analysis of public services in rural... NGP Yearbook 2013 Rantsila 20,12 0,27 4,28 0,049 10,34 0,13 5,50 0,083 0,86 **** 0,74 **** 0,83 0,63 0,82 0,73 0,98 **** 0,86 **** 29.79 km 29.56 km 30.04 km 29.44 km 0,99 *** 0,98 *** 0,99 *** 1*** Kestilä 15,77 0,17 3,71 0,037 7,71 0,076 4,35 0,054 0,67** 0,47** 0,72** 0,48** 0,61** 0,41** 0,77** 0,56** 29.86 km 30.15 km 30.10 km 29.06 km 1*** 1*** 1*** 0,99 *** Piippola 19,62 0,26 4,50 0,057 10,62 0,14 4,50 0,061 0,84 0,72 0,87 **** 0,74 **** 0,84 **** 0,75 **** 0,80 0,64 25.83 km 26.62 km 25.34 km 26.16 km 0,86 ***** 0,88 ***** 0,84 ***** 0,89 ***** Pulkkila 23,45 0,36 5,18 0,078 12,61 0,18 5,61 0,096 1* 1* 1* 1* 1* 1* 1* 1* 22.48 km 23.37 km 21.92 km 22.87 km 0,75 * 0,76 * 0,73 * 0,78 * average distance average distance average distance standardized average distance standardized population relative accessibility 2 Accessibility index population relative accessibility 1 Total Total relative accessibility 1 Youth relative accessibility 2 Youth aged relative accessibility 1Working aged relative accessibility 2Working Elders relative accessibility 1 Elders relative accessibility 2 population relative accessibility 1 standardized Total population relative accessibility 2 standardized Total relative accessibility 1 standardized Youth relative accessibility 2 standardized Youth aged relative accessibility 1 standardized Working aged relative accessibility 2 standardized Working Elders relative accessibility 1 standardized Elders relative accessibility 2 standardized population average distance Total average distance Youth aged Working Elders population average distance standardized Total Youth aged average distance standardized Working Elders Table 1. the results of calculations made in this study. Table * Pulkkila is the most accessible by all measures. ** Kestilä is the least accessible by relative accessibility measures. in average distance values. *** Kestilä and Rantsila have essentially no difference older population, while Piippola the relative accessibility measures, Rantsila is more accessible for **** By is more accessible for younger population. ***** Piippola is the second most accessible by average distance measures.
48 Nordia Geographical Publications 42: 2, 39–52 Mikko Tervo et al. are more extreme. This is a result of the The division of data into three cost of moving along the road network demographic subgroups unearths an and the will to make the trips modeled in interesting difference between Piippola the calculation have a stronger effect. To and Rantsila. By relative accessibility put otherwise, values with α being 2 seek measures, with both α values, Rantsila is to model the accessibility in a situation in more accessible for the elderly population, which movement along a network is costly, while Piippola is more accessible for the difficult or undesired. Average distance youth and the working adult segment values are presented in the same manner in of the population. Also, the advantage table 1, below relative accessibility values. Rantsila has towards the elderly tilts the As can be seen by scrutinizing all the total population accessibility values in measures, Pulkkila is the most accessible of Rantsila’s direction even though Piippola the four municipal centers. This leaves little has much shorter average distance values room for speculation, since the difference for all the demographic subgroups. If the in accessibility is quite large in Pulkkilas’s analyses would have been done only for favour against the other three centers. This the total population, this would not have goes in such extremes, that with α value been noticed. of 2, the accessibility of Pulkkila is more than double the accessibility value of the least accessible center (for example, the Using GIS for accessibility analysis in youth of Kestilä). On the other hand, some a rural setting differences are not very large, for example Based on these empirical results, we argue accessibility values with α of 2, and there that relative accessibility measures can be is hardly any difference between the elders used to present differences in accessibility in of Pulkkila and Rantsila. The general a very simple form and the comparison of superiority of Pulkkila can of course be various locations is easy when this method explained by the central location of Pulkkila is used. However, relative accessibility in the Siikalatva area. The other three measures are somewhat abstract by nature centers however, offer many interesting and a few of the possible caveats should points to address. be mentioned. Firstly, in the method there Kestilä is the least accessible of the is a step in which calculating how many four centers by the relative accessibility people can be reached in a second takes measures even though Kestilä and Rantsila place, which can be seen as being somewhat have essentially no difference in average abstract. Secondly, any accessibility value distance measures. However, the difference does not carry any information when in relative accessibility values between it is cut off from the other values of Kestilä and Rantsila is remarkable. Spatially the relative accessibility matrix. Thirdly, both are located in their respective corners potential accessibility measures used in the of the Siikalatva area, which might be an background are somewhat theoretical and explanation for the similarities in average a valuable question is: Do they even work distance values.
49 Accessibility analysis of public services in rural... NGP Yearbook 2013 on a smaller scale, such as on the scale of a and the different locations of the municipal rural municipality? Caveats aside, we believe centers in relation to the network and the that as a method for aiding local knowledge population. based decision making in combination with other methods, such as average distance measures, relative accessibility is a sound Conclusions method. Average distance is of course in its very We argue that by use of relative accessibility definition an average, a rough estimate and average distance calculation with of what the reality of a resident of the well modeled road network and accurate municipality in question is like. For a person demographic data in GIS environment, this living in a municipal center, the average study has succeeded in proving its usability distance values might seem outrageously in analyzing public service accessibility and high and on the other hand, for the unearthing differences in the accessibility residents living in the extreme periphery of municipal key locations for different the values might be something to dream parts of the demographic in a heavily of. Nevertheless, average distance measures road transport dependant rural setting. do provide a useable estimate of what the On this basis, we argue that these kind of travel distances in the municipality are measures should be taken into account in like, especially for the interest of making situations much like Siikalatva’s merge that comparisons. However, average distance as are taking place in Finland as part of the a method of analyzing accessibility is best municipal restructuring, or in any similar used in unison with other analyses and its processes wherever they might be taking usability lies in the practical nature of the place (and we are of the belief that such values and easy comparisons, as travel costs, reorganizations have been and will be taking be they measured in distance or time used, place around Europe with similar processes are quite easily understood. of outmigration and urbanization), be that It is curious, that in some cases, like the necessary data is available. Our notion Pulkkila for example, relative accessibility here is that the methods built for analyzing values and average distance values are large-scale economic behavior have to be clearly in unison: Pulkkila is relatively applied carefully and with consideration the most accessible center and offers the when operating in on local and rural scale shortest average distances. In some other and variables must be selected carefully, as cases, however, the two analyses might give was done in this study. different outcomes: Average distances for It is to be noted, that while Farrington Kestilä and Rantsila are essentially the same, and Farrington (2005) do give credit to but corresponding relative accessibility empirical measurements of accessibility as values are strongly in Rantsila’s favor. This is a way of gaining some general vision on the result of nuances in the topology of the the (rural) accessibility, they deem them as road network manifested in the accurately normative and imbued with value judgment. modeled data, spatial demographic structure However, we feel that in this paper we have
50 Nordia Geographical Publications 42: 2, 39–52 Mikko Tervo et al. managed to take this into account and view Role of the Funding Source accessibility as a relative and an abstract measure rather than voicing opinions how This paper was made as a part of the public services in rural areas should or Innovative GIS Network in Oulu Region need to be accessed, while still maintaining -project (A30896). The parties involved functionality in decision making processes in the project are University of Oulu, taking place outside the academic world. Oulu University of Applied Sciences, However, as far as decision making Central Ostrobothnia University of Applied processes go, we argue that there is generally Sciences / CENTRIA and the Oulu not enough spatial and demographic Southern Institute. The project is funded information on spatial accessibility available by the European Regional Development to decision makers. We firmly state that GIS Fund, Oulu University of Applied Sciences, analysis should be taken into account more Central Ostrobothnia University of Applied often when making large- scale reformations Sciences / CENTRIA, the City of Oulu, in services. However, accurate georeferenced the City of Ylivieska, the Oulu Southern demographic information is not always Institute and the Kerttu Saalasti foundation. available to decision makers. Also, the role of quantitative information may be avoided in politics, due to its technical and revealing References nature. Spatial accessibility is however not always a straightforward phenomenon Cheng, E. W. L., H. Li, & L. Yu (2007). A GIS that can be grasped without diving into approach to shopping mall location selection. Building and Environment 42: 884–892. demographic analysis in order to reveal Church, R. L. & A.T. Murray (2009). Business possible unseen factors. This can be seen very Site Selection, Location Analysis, and GIS. clearly from the differences in accessibility John Wiley & Sons, New Jersey. for the different parts of the demographic. Clarke, I. & J. Rowley (1995). A case for spatial We are of course certain that accessibility decision-support systems in retail location planning. International Journal of Retail and has many forms and measures with spatial Distribution Management 23: 4–10. network accessibility being only one side Cromley, E. K. & S. L. McLafferty (2002). GIS of the issue. However, we are of the belief and Public Health. The Guilford Press. New that the methods presented in this paper can York & London. Farrington, J. & C. Farrington (2005). Rural be applied to various problems in various accessibility, social inclusion and social locations and scales. The methods excel in justice: towards conceptualization. Journal changing a challenging multidimensional of Transport Geography 13: 1–12. spatial reallocation problem into a form in Finnish Road Administration (2010). Digiroad which it can be grasped without extensive – A National Road and Street Database. http://www.digiroad.fi/en_GB/ (accessed academic knowledge and thus could be of 9.9.2010) broader use in Finnish rural decision making Geertman, S. C. M. & J. R. Ritsema Van Eck – or anywhere else such reorganizations are (1995). GIS and accessibility potential being made. International Journal of Geographical Information Systems 9: 67–80.
51 Accessibility analysis of public services in rural... NGP Yearbook 2013
Geurs, K.T. & J. R. Ritsema van Eck (2001). Ricketts, T. C. (2010). Accessing Health Care. Accessibility measures: review and In A Companion to health and medical applications – Evaluation of accessibility geography. In Brown, T., S. McLafferty & G. impacts of land-use transport scenarios, Moon (eds.): A Companion to health and and related social and economic impacts. medical geography, 521–539. RIVM report 408505 006 Rodrigue, J-P., C. Comtois & B. Slack (2009). Gutiérrez, J. (2001). Location, economic The Geography of Transport Systems. potential and daily accessibility: an Routledge. Abington & New York. analysis of the accessibility impact of Rusanen J., T. Muilu, A. Colpaert & A. the high-speed line Madrid–Barcelona– Naukkarinen (2004). Georeferenced data French border. Journal of transport as a tool for monitoring the concentration of geography 9: 229–242. population in Finland in 1970–1998. Fennia Harris, C. (1954). The Market as a Factor in 182: 2, 129–144. the Localization of Industry in the United Schürmann, C. & A. Talaat (2002). The States. Annals of the Association of European Peripherality Index. Presented at American Geographers 44: 4, 315–348. the 42nd Congress of the European Regional Holl, A. (2007). Twenty years of accessibility Science Association (ESRA). 27–31.9.2002. improvements. The case of the Spanish Spiekermann, K. & M. Wegener (2007). Update motorway building programme. Journal of Selected Potential Accessibility Indicators. of Transport Geography 15: 286–297. Spiekermann & Wegener, Urban and Kotavaara, O., H. Antikainen, M. Marmion Regional Research, (S&W), RRG Spatial & J. Rusanen (2012). Scale in the effect Planning and Geoinformation. ESPON. of accessibility on population change: Spiekermann, K. & H. Aalbu (2004). Nordic GIS and a statistical approach to road, Peripherality in Europe. Nordregio Working air and rail accessibility in Finland, Paper 2004:2. Nordregio – the Nordig 1990–2008. The Geographical Journal Centre for Spatial Development. 178: 4, 366–382. Statistics Finland, 2010. Grid Database.
52