Keywords

1 Introduction

In the last years, because of the increasingly important problem of pollution and high levels of congestion on the road network, many urban centres are affected by sustainable mobility policies and in particular policies which promotes intermodality and pedestrian mobility, undertaken by the Public Administration. It is necessary, in fact, that the public transport supply increases its level of service in order to reduce problems such as high vehicle volumes on the roads, poor attractiveness and low levels of the use of public transport system. Therefore, administrations should design interventions and implement strategies that could guide transport demand towards public transport and sustainable transport modes. Even the city of Palermo is slowly moving towards this direction.

The determination of the accessibility and of the catchment area of the public transport system is a task of fundamental importance for understanding how the transport system will respond to the needs of the potential demand.

Accessibility is defined as “whether or not people can get to services and activities at a reasonable cost, in reasonable time and with reasonable ease” [9].

Therefore, a transport system is widespread and accessible when it is widely distributed on the territory as well as easily reachable from many users as possible, both in spatial and temporal terms. These features make the public transport system dynamic and attractive [16].

The main aim of this work is to evaluate the accessibility of the railway system and the tram system planned by the Administration of Palermo.

2 The Integrated Public Transport System of Palermo

The integrated public transport system of Palermo consists of the rail link, connecting the city to the airport “Falcone e Borsellino” and to the regional rail network, the rail ring, which will be, at the end of the work, a circular line which affects the central districts, and the tram system.

The railway link extends for 30 km, 13 within the city. Currently there are 13 active stations and trains make service with a frequency of two trains per hour and per direction. Interventions provide the doubling of the underground line, the modernization of the existing stations and the building of five new stations, with a total of 18 active stations at the end of the works.

The work on the rail ring provide, instead, its closure, which will be carried out in two stages. There are currently 4 existing stations and other 4 will be realized.

The current tram system consists of 4 lines with 44 active stops. The Administration has provided that the system will be expanded with the creation of 7 new lines, with different priorities, in order to make the transport supply more uniform and widespread in the territory and create a real connected network of trams [21] (Fig. 1).

Fig. 1.
figure 1

The integrated public transport system of the city of Palermo

Full size image

3 Isochrones and Catchment Area

The aim of the methodology developed in this study is to assess whether the planned interventions produce a significant expansion of the catchment area.

The catchment area has been determined by reference to pedestrian accessibility at stops and as the area enclosed by isochrones, which allow you to determine the number of potential destinations that can be reached by pedestrians within a predetermined time, starting from any station.

The isochrones cover distances of 150, 300 and 500 m on foot by the user, necessary to access to the transport network.

Considering a pedestrian speed of 3.6 km per hour, the travel times which characterize the isochrones are shown in Table 1.

Table 1. Isochrones and travel time
Full size table

As shown in Fig. 2, the isochrones can be identified with reference to a topological distance (radial isochrones) or, more accurately, with reference to the metric distance from a point, calculated in reference to a network (spatialized isochrones). In this case, it was preferred to use the first type of isochrones, because we didn’t have a pedestrian network but only a road network for tracing the spatialized isochrones.

Fig. 2.
figure 2

Radial isochrones and spatialized isochrones

Full size image

The catchment area has been determined in terms of resident population and employees potentially served by the system. The data on the resident population in the various census sections provide us indications on the potential origins of the movement, while the data on employees show us the potential destinations.

Demographic data and data on the employees were provided by ISTAT, which has made them during the General Census of Population and Housing and the Census of Industry and Services, both made in 2011 [19]. It is precisely in the identification of catchment area that Geographic Information Systems can play a crucial role [17], being tools that allow you to make spatial operations and associate geographical data with other data type, such as demographic data [2,3,4, 7].

4 Software

For the determination and the visualization of the catchment area, it has chosen to use QGIS, an open source geographic information system, released under a free license. This tool offers a growing number of features provided by the basic functions and numerous downloadable plug in. To create a database in which to import data, called “postgres”, PostgreSQL, an object-relational database, with a free license, has been used. PostgreSQL allows you to query data using SQL. These data are collected in a series of tables with foreign keys, which serve to connect them. In addition, the spatial extension of PostgreSQL, PostGIS, which allows the management of geo-referenced data and confers to the database all the typical spatial analysis capabilities of a GIS, has been installed.

5 Determination of the Catchment Area

In order to visualize tram stops and train stations in QGIS, present and future locations were identified and marked in Google Earth. For each tramline, for the railway link and the railway ring, different place marks, placed in the exact geographical location of stops or stations, were created and were saved as kml files (Fig. 3).

Fig. 3.
figure 3

Placemarks in Google Earth

Full size image

The kml files were then imported as point layers into QGIS clicking “Add Vector”, and the result is shown in Fig. 4.

Fig. 4.
figure 4

Stops and stations in QGIS

Full size image

Using the OpenLayers Plugin of QGIS, the map of the city provided by OpenStreetMap was imported as cartographic base map [20].

A second phase involved the visualization of the data on population and employees of the municipal area in GIS. We referred to the data on the Census made by Istat in 2011. ISTAT, on its website, publishes the territorial bases and geographic data of each census section, which is located within the municipal area; these geographical data can be viewed through a GIS software since they are given in shp format. The Institute of Statistics also provides the data of the General Censuses of Population and Housing and of the Censuses of Industry and Services in xls format; through QGIS and the PostgreSQL database, it has been possible to associate the census data to the census section [18].

The shp file of census sections, downloaded from the website of ISTAT, has been loaded as a new project layer. The result is shown in Fig. 5.

Fig. 5.
figure 5

Census sections

Full size image

The reference system is WGS 84 UTM Zone 32 N. The municipal area has been divided into 2834 census sections, with different extensions. There is a condensation of the census sections in the central districts of the city, while the suburbs are divided into larger census sections.

The shape file was also imported into the database “postgres” as table called “r19_11_wgs84”, through the application PostGIS Shapefile Import/Export Manager. Then census data, collected in an Excel file, were downloaded from ISTAT website. In the “postgres” database it was created a table, called “population”, and the empty columns “sec”, in order to receive the identification numbers of the census sections, and “population”, in order to receive the corresponding values of the resident population, were added.

The following strings in SQL have been written:

figure a

The columns were filled with data from the Excel file, relative to the census sections and values of population residing in them, saved at first as a text file in .txt format. The creation of a new table, called “census_section”, in the database and the mix between the table “population” and the table “r19_11_wgs84,” has made it possible to have geographic data and population data in a single vector. The query was as follows:

figure b

This vector has been imported as a new layer in QGIS using the command “Add Vector PostGIS” and its visual style has been modified so that census sections are coloured with different shades of colour depending on the value of the population living in them (Fig. 6).

Fig. 6.
figure 6

Population

Full size image

Similar operations have been carried out for the data on employees. ISTAT, in the realization of the Census of Industry and Services, has divided the employees, depending on the business sector, in:

  • employees in the industrial sector;

  • employees in public institutions;

  • employees in no profit institutions.

Thanks to the QGIS plugin “mmqgis” has been possible to make buffering operations around the stations and identify the isochrones corresponding to distances of 150, 300 and 500 m walked by a pedestrian [10,11,12] (Fig. 7).

Fig. 7.
figure 7

Buffering

Full size image

However, the identified catchment area does not give the number of residents and employees of different sectors, who are intercepted by the transport system. In order to achieve this result, the intersection between the layer of isochrones and the layer of census sections, which has the data on population and employees as attributes, has been made in QGIS, clicking on “Vector/Geoprocessing tools/Intersection”.

Some census sections are partially within the catchment area. In order to determine the effective population/number of employees that is in the catchment area further steps are necessary. Using QGIS it is possible to evaluate the effective area of each census section that is within the catchment area and the effective population or number of employees that each census section owns.

In the attribute table it is possible to click on “field calculator” adding the three columns “areaeff”, “perc.area” and “popeff” (or “addimpeff”, “addipeff” and “addnpeff”).

An approximation about the distribution of population and employees in the census sections has been made. In fact, the population was regarded as uniformly distributed within the census section. With this simplification, the effective population as a product of population of the entire census section area for the percentage of area of the census section which actually falls within the catchment area has been calculated.

These modifications have been saved and the layers have been saved with a name. For example, the layer regarding the population which falls in the catchment area of 500 m of the integrated system, is saved with the name of “pop_sistema_integrato_500”, while that relating to employees in the industry with the name of “add_imp_sistema_integrato_500”. The shape files, which were created for each considered scenario and for the fixed distances of buffering, have been imported into the PostgreSQL database with the application “PostGIS Shapefile Import/Export Manager”.

Finally, for each scenario, the total population and the total number of employees, which fall in the catchment area generated by distances of buffering of 150, 300 and 500 m, are calculated by means of queries in SQL language similar to the following:

figure c

The next paragraph shows the results in table form (Tables 2, 3 and 4) and the display of the catchment areas in QGIS for each scenario.

Table 2. Potential demand (150 m)
Full size table
Table 3. Potential demand (300 m)
Full size table
Table 4. Potential demand (500 m)
Full size table

6 The Decision Support System and the Design Scenarios

Three scenarios have been considered. Scenario 1 (Fig. 8) is representative of the current public transport system, because there are the stations of the railway link and the stops of the tram lines currently operating (lines 1, 2, 3 and 4). Scenario 2 (Fig. 9) takes into account the addition of the rail ring, while the scenario 3 (Fig. 10) contemplates the extension of the tram system with the addition of new lines.

Fig. 8.
figure 8

Scenario 1

Full size image
Fig. 9.
figure 9

Scenario 2

Full size image
Fig. 10.
figure 10

Scenario 3

Full size image

The mentioned analyses, which are developed in a GIS environment, allow the development of a decision support system able to calculate the potential demand for the integrated transport system, varying the axes that constitute the public transport network in the design phase. The transport demand may be computed using the joint probability that the origin-destination pair of any journey falls within the area of influence of the stations/stops of the integrated system.

As mentioned before, the resident population and the employees of individual census sections are proxy variables of the origins and destinations of the journeys for different reasons for the journeys.

This analysis allowed the identification of interventions that involve the most significant benefits and of those which had less significant effects.

7 Discussion

The results show that the realization of the new tram lines could intercept a larger number of users. In fact, considering a radius of the catchment area of 300 m, which is the most frequently considered distance for this kind of analysis, around 36% of the transport demand is potentially intercepted for the scenario 3, while only 7% of the demand is potentially intercepted by the transport supply in scenario 2. Considering a radius of 500 m, the integrated system is able to potentially cover more than 50% of total O/D pairs and, in particular, the closure of the rail ring and the expansion of the tram system could be crucial in order to intercept a larger catchment.

Some tram lines are uneconomic, i.e. lines E and F, since they have high investment costs due to their length but they involve very small increments of the transport demand intercepted.

Once the potential demand will be determined, the use of an appropriate random utility model for evaluating the demand modal split will be able to support the analyst to calculate the reliable demand attracted by the integrated transport system [1, 15].

A cost-benefit and a multicriteria analysis will assess the success of individual lines within the network at the design stage, taking into account, on the one hand, of the benefit of upgrading infrastructure computed as a function of the transport demand satisfaction and of the services offered from the transport system, and, second, of the costs of the infrastructure investment and its management, as well as of the positive and negative externalities generated by the investment [5, 8, 14].

In this case study, for the evaluation of the reliable transport demand in the different scenarios, the modal split has been calculated using a multinomial logit model, considering the service of the railway link, the railway ring, the tram lines and some bus lines with an high level of frequency, taking into account the introduction of the Limited Traffic Zone and the parking pricing policy [13]. The modal split, as result of this analysis, has been equal to 40% of the potential demand [6].

So, considering a radius of buffering of 500 m, the reliable demand attracted by the integrated system has been evaluated as the 40% of the potential demand and the results are shown in Table 5. Therefore the integrated transport system designed for the city of Palermo could be able to intercept the 20% of the transport demand.

Table 5. Reliable demand (500 m)
Full size table

8 Conclusion

The adopted methodology has been developed to evaluate the spatial accessibility and the social equity of an urban public transport system. Currently, it has been applied to the public transport network of the city of Palermo in order to analyze the spatial accessibility and the equality in the distribution of urban services and the impact that the planned transport projects will have on spatial accessibility by public transport.

Therefore, it compares the zone accessibility by public transport and can estimate the accessibility impacts by proposed transport infrastructure changes.

The methodology also provides an overview of the attractiveness of the zones in order to identify the “hotspots”, which are areas of potential congestion that may require specific management approaches.

It can also identify those zones that are relatively poorly served by public transport system.

Moreover, the study was conducted to estimate the transport demand, to develop procedures that can be standardized and systematically applied, and, finally, to investigate the travel behaviour of the resident population.

The methodology has been designed and developed not only to estimate the demand within the City of Palermo, but in perspective of an application in similar territorial context or also in a different one. For an example, the methodology could be used with other design scenarios containing new distributions of population and employees in the various zones using models that predict the distribution of the residents on the territory as function of new levels of accessibility. In fact, the strength of this methodology is the ability to take into account both transport and land-use systems for accessibility analysis and for the evaluation of the transport demand.

Then this work showed how GIS is an effective decision support system for the Public Administration, allowing it to identify the most effective project proposals. Furthermore, this methodology has allowed the construction of the catchment areas and the determination of the demand potentially served by a public transport system without resorting to the construction of an O/D matrix, which is always approximate; we used accurate data such as those on the resident population and on employees.

The use of spatialized isochrones, generated on the basis of a pedestrian network, would have involved a larger accuracy in the determination of the results.

The adopted methodology allowed us to verify that the scenario designed by the Municipal Administration actually manages to ensure a significant expansion of the catchment areas and makes the integrated system of public transport rail more attractive in transport demand, compared to the current transport system.

For the realization of an integrated system of this type much time and money will have to be employed and the administration’s efforts will be rewarded only through proper management of the system, especially in terms of choice of frequencies and choice of service fees. These choices will be critical to ensure that the system will attract a significant part of the transport demand from the private car.