Trento is a medium-sized city of around 120,000 inhabitants located along the course of the River Adige, in the eastern Italian Alps. The intensely-urbanised valley floor hosts around 70% of the population, as well as most of the industrial areas, commercial units,and transport infrastructures (Fig. 1). The rest of the population live in small villages on the hills, surrounded by agricultural areas, predominantly vineyards and apple orchards. The large municipal territory (around 156 km², half of which covered by forests) includes the surrounding mountains up to an elevation of 2,180 m. Natural protected areas account for more than 10 km², including seven Natura 2000 sites. The presence of close-by natural areas is a peculiar feature of the city, which affects the type of demand and uses of urban green infrastructure. However, not all neighbourhoods in the city equally benefit from green spaces: the northern suburbs, which host a population mostly composed of low-income families, are the most disadvantaged areas, characterised by an inadequately-planned mixture of industrial, commercial, residential and agricultural uses; disconnection from the city centre; and a scarce presence of services and public spaces. Other key challenges currently faced by the city, notably affecting the urban areas in the valley floor, are the consequences of extreme climate events, especially heatwaves and urban flooding.

Figure 1.  

Main land uses in Trento, Natura 2000 sites and the 13 brownfields identified by the urban plan as ‘urban redevelopment sites’. Source: Comune di Trento (2017a).

The presence of brownfields in the most dense and populated part of the city is one of the main planning issues in Trento. Amongst the existing brownfields, the current urban plan identifies 13 areas called ‘urban redevelopment sites’, whose regeneration is considered a priority to prevent or counteract the emergence of social, economic and environmental problems. The 13 sites range in size from 0.5 to 9.9 ha and cover a total area of 44 ha (Fig. 1). Until now, the costs - especially in the case of contaminated sites - and the bureaucratic burden associated with interventions, as well as the sometimes contrasting interests of public administration and private owners, have hindered their regeneration. Given these premises, it is reasonable to assume that only some of the brownfields will be converted to new industrial or residential areas in the next years. A greening intervention can thus be advanced as a possible, perhaps temporary, solution. Building on Geneletti et al. (2016), this study is aimed at supporting the decision about which of the existing brownfields should be converted into a new urban park based on the benefits produced for the surrounding inhabitants.

Two key urban ES for Trento are considered in the assessment of brownfield regeneration scenarios, namely microclimate regulation and recreation. The selection of microclimate regulation is linked to the growing concerns of the local administration for summer heatwaves, particularly intense in the city due to the low altitude and the narrow nature of the valley. As demonstrated by the 2003 event, Trento is more vulnerable to heatwaves compared to other Italian cities (Conti et al. 2005). Heatwave effects combine with the urban heat island in the valley floor, the most urbanised and sealed part of the city (Giovannini et al. 2011), causing peaks in energy demand for cooling (Grimmond 2007) and posing serious threats to citizens’ health and well-being (Kenny et al. 2009). Considering the increased frequency and intensity of heatwaves expected in the coming decades (Fischer and Schär 2010), effective solutions to mitigate their negative effects and to provide cool areas for heat relief during the hot season are seen as one of the most pressing needs by citizens and administration.

The focus on recreation is in line with the main planning objectives of the city administration. In the last years, the enhancement of public green areas has been targeted toward gaining a more balanced distribution over the city, hence providing equal opportunities to all citizens for recreation and relaxation. However, understanding these opportunities for nature-based recreation, i.e. outdoor recreational activities that imply physical or experiential interactions with natural components of the environment (Haines-Young and Potschin 2018, Zulian et al. 2013) being equally distributed, is not an easy task. In Trento, besides urban parks, citizens also benefit from the proximity to other typologies of green areas where they conduct a wide range of activities, including hiking, mountain-biking, skyrunning and climbing. Indicators based on the availability of and accessibility to urban parks are not enough to capture this variety (Paracchini et al. 2014). Assessing recreation as an ecosystem service, considering different providing units and different levels of demand, could provide planners with useful information for achieving an equal distribution of recreational opportunities over the city (Kabisch and Haase 2014).

The regeneration of existing brownfields through their conversion to new urban parks is an opportunity to enhance several key ES for urban areas (Collier 2014, McPhearson et al. 2013), including the two ES selected for the study. Urban green areas, especially when carefully designed (e.g. shape, vegetation density, presence of water), not only reach an internal temperature significantly lower than built-up areas, but can also produce a cooling effect on the surroundings (Zardo et al. 2017). From this point of view, the location of the thirteen 'urban redevelopment sites' in the valley floor (Fig. 1) represents an opportunity to mitigate the combined effects of heatwaves and the urban heat island. At the same time, the regeneration of these brownfields strategically located close to the city centre can provide new recreational opportunities for citizens living where both the availability of urban parks and the proximity to natural areas are low (De Sousa 2004).

To map and assess the cooling effect of urban green infrastructure, we adopted a method specifically designed to support planning and management decisions at the urban and sub-urban scale. The method is described in Zardo et al. (2017) and Fig. 2 provides an overview of the main stages. The method accounts for the two main ecosystem functions involved in microclimate regulation, i.e. shading and evapotranspiration and assesses them based on three properties of green infrastructure components, namely soil cover, canopy coverage and size. Once the green infrastructure component is classified according to the three properties, the model provides the corresponding cooling capacity score, depending on the climatic region of the study area (see Zardo et al. (2017)and Fig. 2 for further details on the model). Then, the cooling capacity scores can be classified and each class can be linked to an expected temperature difference between the analysed area and a reference area with the lowest cooling capacity (i.e. a sealed surface with no trees). For the present application, we adopted a version of the model with the range of scores scaled up to a maximum value of 172 and six final classes of cooling capacity, from A+ to E, as in Geneletti et al. (2016). Finally, the cooling effect produced on the surroundings can be mapped using decay functions. The model accounts for the effect of evapotranspiration by assuming linear decay functions that vary depending on the size of the area, whilst the effect of shading is approximated by a local buffer around canopies. The cooling effect can be mapped using the same classes defined for the cooling capacity, with the same range of expected temperature difference (Geneletti et al. 2016).

Figure 2.  

Flow chart of the model for mapping and assessment of the cooling capacity and cooling effect of urban green infrastructure, building on Zardo et al. (2017). Input data are listed on the left. For model parameters (i.e. scoring tables and distance decay functions) refer to Geneletti et al. (2016), Zardo et al. (2017).

To assess the improvement in micro-climate regulation under the planning scenarios, the new urban parks obtained by the regeneration of existing brownfields were modelled as areas covered by grass, with 80% to 100% canopy coverage. Maps of the cooling effect were produced for the baseline condition and considering the conversion of each brownfield (i.e. each scenario) independently. Then, we computed the difference between each scenario and the baseline condition and intersected the resulting maps with a map of population distribution. The final indicator for each scenario was defined as the number of affected residents weighted by the intensity of change in the class of the cooling effect of their location (i.e. residents experiencing an improvement of two classes are counted twice). Young children (<5 years) and the elderly (>65 years) were selected as the most vulnerable groups, based on their higher sensitivity to heat stress (Basu 2002, Kabisch et al. 2017, Kenny et al. 2009) and counted separately.

To map and assess the potential and opportunities for nature-based recreation in the city, we applied a locally-adjusted version of ESTIMAP-recreation. The model is part of the ESTIMAP suite, originally developed by the European Commission's Joint Research Centre for the purpose of mapping ecosystem services at the European scale (Zulian et al. 2013, Paracchini et al. 2014) and later adjusted for the application to different contexts and scales (Zulian et al. 2018). The model is composed of two modules, as summarised in Fig. 3. The first module assesses the Recreation Potential (RP), i.e. the suitability of different areas to support nature-based recreational activities based on their intrinsic characteristics and produces a raster map with relative values ranging from 0 (no recreation potential) to 1 (maximum recreation potential in the analysed area). The second module assesses the Recreation Opportunity Spectrum (ROS) by combining the RP with information about the availability of infrastructures and facilities to access (e.g.cycle paths, bus routes, parking areas) and to use (e.g. playgrounds, sport fields, park furniture) the area, thus providing an assessment of the opportunities offered to the citizens. The module produces a raster map classified into 9 categories resulting from the cross-tabulation of high/medium/low RP and high/medium/low availability of infrastructure and facilities. The different elements that contribute to the values of RP and ROS, listed in Fig. 3, are combined according to scores assigned by the user. For the described application, the scores were elicited from a pool of local experts, as detailed in the 'Data' section.

Figure 3.  

Flow chart of the ESTIMAP-recreation model adjusted for the application to Trento. Modified after Zulian et al. 2013.

To assess the enhanced opportunities for nature-based recreation under the planning scenarios, the new urban parks obtained by the regeneration of existing brownfields were assigned to the land use class ‘green urban areas’ and assumed to be equipped with the same infrastructure and facilities as other parks with comparable dimensions. People living within 300 m from the new parks were considered as the main beneficiaries of the transformation (Kabisch et al. 2016, Stessens et al. 2017). Children and teenagers (<20 years) and the elderly (>65 years) were selected as the most vulnerable groups, based on the higher demand for close-to-home recreation and relaxation areas (Kabisch and Haase 2014). Furthermore, we distinguished the beneficiaries already served by high-level opportunities for nature-based recreation in the baseline scenario (i.e. living within 300 m from areas classified in the highest class of ROS) and counted them separately.

A multi-criteria analysis was used to combine the results of the two ES assessments. The 13 scenarios simulating the regeneration of the different brownfields were considered as alternatives. The two ES and the different categories of beneficiaries based on the level of vulnerability were used as criteria and sub-criteria, respectively (Table 1). We applied three illustrative combinations of weights, corresponding to three hypothetical policy perspectives and related objectives (Table 1). The 'cool air for the elderly' perspective favours improvement in the cooling effect in areas with a high share of older population. The 'every child needs a park' perspective favours opportunities for recreation to people, especially children and teenagers, who are not served in the present condition. The 'balanced' perspective promotes both ES equally, but gives more weight to the most vulnerable beneficiaries (see Table 1 for details on the weights). Values for each criterion and sub-criterion were normalised according to the maximum and a ‘weighted summation’ approach was used to calculate the overall score for each alternative, hence defining the final rankings for the three perspectives. Finally, a sensitivity analysis was conducted to explore the robustness of the rankings to variations in the weights assigned to criteria and sub-criteria.

Information about green infrastructure in Trento were mostly retrieved from municipal data, including a land use map published in 2017 (Comune di Trento 2017a) and the municipal database for the management of public green areas and trees (Comune di Trento 2010), which provides detailed information about their typology and location (including data about tree species, age and dimension) and about the presence of facilities in urban parks (e.g. benches, fountains, playgrounds). To assess the micro-climate regulation, first the land use map was integrated with more detailed information available from other sources for specific areas (e.g. municipal database of public green areas, map of community gardens). Then, each land use class was assigned to one of the five soil cover classes identified by the model (Table 2). Canopy coverage was mapped by combining the land use map with the provincial and municipal maps of forested areas and the municipal database of public green areas and trees.

Input data for the ESTIMAP-recreation model were retrieved from both institutional databases and Open Street Map (Open Street Map Contributors 2017). The scores were elicited through an on-line questionnaire from a pool of experts selected with the collaboration of a municipal officer. The questionnaire was sent to 19 experts who had previously agreed to collaborate to the project and 17 valid responses were collected within the deadline (December 2017). Respondents include personnel of different provincial (3) and municipal (7) departments with an interest in recreational areas and activities, local practitioners (1) and academics from the University of Trento (3) and other research centres in the city (3). The experts were asked to assign to each element a score from 0 to 5 based on the role of the element in supporting or promoting nature-based recreational activities. The scores were then averaged, excluding the highest and the lowest scores and converted to a 0-to-1 scale. The final scores used to run the model are listed in Table 2 and Table 3.

Population data for each census tract, including 5-year age groups, were also provided by the municipality (last update: 31^st December 2014). To be as accurate as possible in the analysis, the population in each census tract was distributed only on the surface covered by the footprint of residential buildings. Spatial data were analysed and elaborated using the GIS software QGIS 2.18.9 (QGIS Development Team 2017) and GRASS GIS 7.2.1 (GRASS Development Team 2017), while the multi-criteria analysis was conducted using the free version of the software Definite (SPINlab Vrije Universiteit Amsterdam 2016).

The assessment of the cooling effect produced by green infrastructure in Trento was carried out for the most urbanised area of the city, i.e. the valley floor, where all the brownfields are located (Fig. 4). Overall, the highest classes of cooling effect prevail there, due to the presence of close-by forests and of the River Adige and its tributaries that contribute to lower the temperature of the surroundings. The most disadvantaged areas are in the dense neighbourhoods close to the city centre and in the northern suburbs. Here, the mix of residential and industrial areas, with little green infrastructure, as well as the high rate of soil sealing generated by the concentration of major transport infrastructures, have a negative impact on the cooling performance of the city. Most of the brownfields appear to be strategically located in areas that scarcely benefit from the cooling effect of both urban green infrastructure and the surrounding natural and semi-natural areas.

Figure 4.  

Map of the cooling effect of urban green infrastructure in the most urbanised part of the city of Trento (baseline condition) and an example of a planning scenario related to the regeneration of brownfield 11. The zoom shows the maximum distance potentially reached by the cooling effect generated by the converted brownfield.

An example of how the conversion of brownfields into new urban parks would affect the cooling effect is provided in the right side of Fig. 4 for the case of brownfield 11. Due to the change in the soil cover from partly sealed to grass and to the intense plantation, the site would reach the highest class of cooling effect. The immediate areas would also be affected by the transformation. In the present condition, most of the surrounding residents gain very little or no thermal benefit at all from the presence of green infrastructure, which is almost exclusively limited to single shading trees. In the regeneration scenario, an improvement is noticeable in the major part of the area. In the neighbourhood to the north, some households would shift from class E to class A of the cooling effect.

The performance of the different scenarios in terms of microclimate regulation is summarised in Fig. 5, where brownfields are compared according to the number of people that would benefit from their regeneration. Brownfield 11, a potentially large green area inside a densely built-up and populated part of the city, is by far the best performing one: more than 2,000 citizens would benefit from the enhanced cooling effect produced by the transformation. The regeneration of the other brownfields is expected to affect much less people, within the range of some hundreds for most scenarios.

Figure 5.  

Expected benefits produced by the different scenarios in terms of enhanced cooling effect by urban green infrastructure: number of beneficiaries broken down into different vulnerability classes.

The map of the Recreation Opportunity Spectrum (ROS) in the city of Trento, as obtained from a cross-tabulation between Recreation Potential (RP) and the level of availability of infrastructures and facilities, is shown in Fig. 6. The valley floor, though mostly characterised by low values of RP, presents the highest concentration of infrastructure and facilities that provide access and support the use of green areas. The best-performing areas in this part of the city are urban parks and the river banks, which host an important touristic cycle path, intensely used by Trento citizens for running, cycling and skating. In the extra-urban areas, different opportunities for nature-based recreation characterise the two sides of the valley. Forests on the east side are characterised by a higher density of forest tracks, hiking trails and facilities dedicated to specific activities such as climbing routes and MTB trails, especially near the settlements. On the west side, the settlements are more sparse and the connections with the valley floor are more difficult, which determines a lower availability of infrastructure and facilities.

Figure 6.  

Map of the Recreation Opportunity Spectrum (ROS) in Trento calculated through the locally-adjusted version of the ESTIMAP-recreation model (baseline condition) and example of planning scenario related to the regeneration of brownfield 11. The zoom shows the 300-m buffer used to identify potential beneficiaries of enhanced close-to-home recreational opportunities.

Considering the brownfields and their surroundings, all of them are in areas with high availability of infrastructure and facilities. Some are close to existing urban parks, as in the case of brownfield 10, while others, e.g. brownfields 01, 02 and 03, are far from any area of high RP. Hence, they represent opportunities to enhance the condition of people that currently have no or very few close-to-home opportunities for nature-based recreation.

Regeneration interventions would convert existing brownfields into new urban parks, thus increasing the opportunities for nature-based recreation in the neighbourhoods. Fig. 6 shows the case of brownfield 11 that, once regenerated, would fall into the best class of ROS, with high RP and high availability of infrastructure and facilities. The map also highlights the possibility of connecting the new park to an adjacent open-air soccer field, already classified in the best class of ROS. Despite being already served by other parks and green areas close by, all the households, included in the map, would benefit from an additional space for recreation within walking distance from their location.

The performance of the different scenarios in terms of recreation opportunities is summarised in Fig. 7, where brownfields are compared according to the number of people that would benefit from their regeneration. Brownfields 07 and 08 produce the highest absolute number of beneficiaries. However, only the scenarios that consider the regeneration of brownfields 01, 02 and 03 would serve people that, at present, have no access to close-to-home nature-based recreational opportunities. The ratio between total beneficiaries and specific vulnerable groups is not the same across scenarios, due to the uneven distribution of population groups across the city. For example, the share of children and teenagers is higher for scenarios 01 and 02 compared to the others, while the share of people aged more than 65 is the highest for scenario 11.

Figure 7.  

Expected benefits produced by the different scenarios in terms of enhanced opportunities for nature-based recreation: number of beneficiaries broken down into different vulnerability classes.

By comparing Fig. 7 with Fig. 5, the performance across scenarios appears more balanced in the case of recreation than in the case of microclimate regulation. Moreover, excluding scenario 11, the number of people that would benefit from increased opportunities for nature-based recreation is much higher compared to the number of citizens that would experience an improved cooling effect. Although some of the beneficiaries overlap between the two ES, due to the spatial relationship between providing units and benefitting areas, the information is important for defining the weights of the different criteria in the multi-criteria analysis.

The information about the number of beneficiaries of the two ES in the different scenarios was combined through a multi-criteria analysis according to the three perspectives described in Table 1. The results of the analysis are summarised in Fig. 8. When assuming a ‘balanced’ perspective, with the same weight assigned to the two ES and a double weight assigned to vulnerable compared to non-vulnerable groups, brownfield 11 ranks first. The second perspective, consistent with the objective of improving the cooling effect in areas with a high share of older population, leads to the same first-ranking scenario. Although the other positions change between the two perspectives, all scenarios gain a very low score compared to brownfield 11. Under the third perspective, the final ranking changes significantly and the first positions are occupied by the three brownfields (01, 02 and 03) located in the northern part of the city. In such neighbourhoods, the population is comparatively younger and the opportunities for recreation are scarcer.

Figure 8.  

Final rankings of the regeneration scenarios according to three perspectives considered in the multi-criteria analysis. The weights assigned to the different ES and the different categories of beneficiaries are reported in Table 1. Scenarios 04, 09and 10 were identified as sub-optimal alternatives and excluded from the multi-criteria analysis.

Overall, the three illustrative perspectives show how priorities for brownfield regeneration change based on the relative importance attributed to the different ES and the respective categories of beneficiaries (Fig. 9). To assess the stability of the ranking and the robustness of the results with respect to possible variations in the assigned weights, we conducted a sensitivity analysis. Considering the weights assigned to the main criteria, i.e. the two ES, the rakings produced by perspective 1 ('balanced') and 2 ('cool air for the elderly') are very stable: the first-ranking alternative (brownfield 11) maintains its position even for large fluctuations of the weight (up to the weight of the ‘recreation’ criterion equal to 0.85). In the case of perspective 3 ('every child needs a park'), the ranking is less stable and a weight of 0.58 assigned to the ‘recreation’ criterion is sufficient for scenario 02 to reach the first position.

Figure 9.  

Map of the priority level of brownfield regeneration scenarios according to three perspectives considered in the multi-criteria analysis.