*4.2. The Provision and Distribution of Urban Parks with Respect to the Population*

The surface area of the urban parks in relation to the inhabitants shows an average available surface area of 2.62 sqm. per inhabitant. This figure rises to 15.09 m2/inhab. with all the green spaces (periurban) in the municipality. The value achieves the thresholds recommended by the World Health Organization, which establishes a figure of around 10–15 sqm./inhabitant. Regarding their accessibility, 39% of the population in the Tarragona capital live within 300 m of an urban park and the average distance in a straight line is 710 m, which is much higher than recommended (Table 2).

If the urban areas are used as the analysis unit, we can observe that the provision of urban park surface area per inhabitant is unequal according to the area of residence. This geographic distribution shows an important degree of environmental injustice. Therefore, some areas appear without any urban parks inside them (the historical center and Sant Salvador), whereby the rate is 0 sqm./inhabitat. In some areas, such as in the case of Bonavista, the parks lie in the periphery of the urban area. In others, such as Torreforta and Ensanche, there is a minimum provision per inhabitant, resulting from the combination of its relatively high population and the low surface area of the existing parks. The area with the residential estates in the eastern area (Levante) has below-average values, but, due to its urban characteristics, it is marked by a low density, the presence of private green areas and its disconnection from the city.

#### *4.3. The Quality of Urban Parks*

The average PQI in Tarragona is 53.39 points, with the highest assessment occurring in vegetation (67.86), followed by street furnishings (45.50) and facilities (32.32) (Table 3 and Figure 6). Out of the 14 urban parks analyzed, none of them showed a "good" quality (PQI equal to or higher than 70 points) and only four have a "medium-high" rating (60–69 points). Specifically, these are the Puig i Valls park (62.81), the Riu Francolí park (62.22), the Amfiteatre park (61.42) and the Fitolaca park (60.29). Their overall score (total assessment of vegetation, facilities and furnishings) is due to different factors. Therefore, the Puig i Valls and Amfiteatre parks owe their high assessment to the vegetation (number of examples, green shadow and absence of allergenic species, 80.00 and 83.33, respectively), although they have lower scores for the other two indicators (facilities 39.00 and 32.08; furnishings 40.02 and 59.15, in the same previous order). On the contrary, the good score obtained by the Riu Francolí park is not due to the assessment of its vegetation (with 53.33 points, it is the third lowest urban green area in the city). Instead, it is due to the good quality of both its facilities (74.78) and its furnishings (67.43), in which it leads the city ranking as a whole.

**Table 2.** Average surface area available per inhabitant and average distance to the nearest urban park.


Source: own work.


Source: own work.

**Figure 6.** Location and quality level of the urban parks in Tarragona. Source: own work.

For the "medium-low" score (between the values 50 and 59 in the PQI), we find four parks: Bonavista, Sant Pere i Sant Pau, la Col·lectiva and el Camp de Mart. The good quality of their vegetation, which helps to offset their fairly mediocre scores in the other two PQI components, gives them an intermediary score. The most representative case in this respect is the Col·lectiva park, where the quality of the vegetation (80 points, the sixth highest) offsets the discrete score for its facilities (15.65).

Finally, we found six parks with a "low" assessment for their quality (PQI under 50 points): Saavedra, Sant Antoni, La Reconciliació, plaça de les Amèriques, Camí de la Coma and Rodríguez de la Fuente. In all of them, with the exception of the Camí de la Coma park, the best scoring indicator was for the vegetation, while, in the other two indicators (facilities and furnishings), low values were obtained (Table 3). The extreme cases are the furnishings score in the Rodríguez de la Fuente park (16.55) and, particularly, the score for the virtually non-existent facilities in the Sant Antoni park (1.25).

The results of the analysis reveal a meagre relationship between the location of the parks in the urban area and their quality. In terms of environmental justice, it could be expected that the parks located in the more peripheral areas were of a lower quality, while those in the center of the urban hub were better quality. However, in all the urban areas, there are parks with different assessments according to their PQI. To provide just one example in this case: two urban parks very close together, the Amfiteatre park and Reconciliació park, separated by just one street, have very different PQI values: while the first one has "medium-high" quality (61.42 points), the second one has "low" quality (46.66).

## *4.4. Environmental Justice Regarding Socio-Economic Characteristics: The Level of Studies, the Place of Birth and Housing Prices*

As we can see from Figure 7a, the relationship between the PQI value and the standardized STI of the population living within 300 m reveals a degree of environmental inequity. On the one hand, with the positive value of the R<sup>2</sup> coefficient and, on the other hand, the actual positive slope of the trend line, it can be concluded that there is a causal relationship between the parks with a lower level of quality and the lower level of studies among the population living within 300 m. In spite of this, this relationship is not particularly robust, with an R2 coefficient value of 0.1159. One factor that influences this behavior lies in the lower values of some urban parks, which are much lower than the other analyzed cases. In the urban parks with better values, the relationship between their

coefficient is not so clear; in other words, the higher values in a variable correspond to the highest scores in the other one.

(c) (d) **Figure 7.** Correlation between the quality of the parks (PQI) and the socio-demographic variables of the population living within 300 m of the nearest park: (**a**) study level index for the population; (**b**) average value of HDI; (**c**) Average value of the purchasing price (euros/m2) and (**d**) average value of the rental price (euros/m2). Source: own work.

The results of the correlation between the HDI of the place of birth and the PQI reveal a clear correlation: the PQI value for each park and the average value of the HDI correlate with an R<sup>2</sup> of 0.1667 (Figure 7b).

The correlations between the PQI and the average price per square meter that is purchased or rented (Figure 7c,d, respectively) show fairly similar situations: if the dwelling is purchased or rented, its price is higher among those located near the greater quality parks. There is also a clearer relationship (R<sup>2</sup> = 0.2123) in the case of purchased dwellings, as the result of a direct linear correlation, whereby the lowest values of one variable are also related to the lowest indicators of the other variable and vice versa. It could be presumed that this direct relationship has something to do with the more or less central location of each park with respect to the city as a whole, understanding that the housing prices follow a more or less concentric logic, where the more central apartments have higher prices than those located in peripheral areas.

At any event, although this idea in the case of the dwellings is certain, it has already been mentioned that the quality of the parks does not follow this pattern, and so the robustness of the relationship is not due to this trend. A weaker, but equally positive (Figure 7c,d) relationship is the one between the price of home rentals and each park's PQI (R<sup>2</sup> = 0.1530).

#### **5. Discussion**

This article assesses the environmental justice in the city of Tarragona (Spain) with respect to the accessibility, availability and quality of the urban parks and the sociodemographic characteristics of its population. According to the Regional Office for Europe of the World Health Organization [22], the minimum surface area of urban green areas must be between 10 and 15 square meters per inhabitant, and within a distance of 300 m or 5 min walk from the dwelling. However, some authors [41] maintain that this strictly quantitative measurement of the provision of green spaces is not enough, and that the

parks must fulfil three basic conditions: availability (they are within a distance that allows its potential users to enjoy it), accessibility (when the user feels welcome, can access the park freely and use it for recreational purposes at any time) and, finally, a certain degree of attractiveness (when the space responds to individual needs, expectations and preferences).

A significant part of the work on environmental justice and urban parks has focused on assessing the distribution of green areas in the city [42,43] by calculating the distance between the place of residence and the nearest green area and using a geographic information system [44]. Even though the most frequently used unit is the distance in meters, some authors choose to measure accessibility using the travelling time according to the means of transport used (public transport, by foot, by bicycle and private car) [45]. Calculating these indicators using Euclidean and network distances shows a clear influence of the type of distance chosen (Euclidean versus the distance in the network). Therefore, we have to use these indicators carefully as planning support tools.

Analysis of accessibility, understood to be the physical or time separation between the actual location of the park and the users' place of residence, is complemented by architectural accessibility issues (for example, if it is a non-closed space); psychological accessibility (if it is attractive enough for potential users to visit) [41] or its "walkability", i.e., whether parks are accessible to people with limited mobility, such as children or the older population [46].

Some of the limitations found in this study are related to urban mobility. By selecting a population within a certain distance threshold with respect to a park, you start with the premise that the population only uses the urban parks in their own urban residency area, in other words, as if these urban divisions led to "islands" or self-contained compartments, without people moving from one to the other. Obviously, this is not entirely true, since mobility is a fundamental component of cities, conditioned by the place of residence and the locations that people visit regularly (work, leisure areas, daily shopping, children's study area). Therefore, beyond the analysis of the provision of urban parks based strictly on the place of residence, it would be important to note the population's daily mobility. The type of mobility used here is by foot, following the recommendations of the World Health Organization. However, due to the promotion of public transport as part of countries' commitment to reduce the consumption of hydrocarbon fuels, and the proliferation of Personal Mobility Vehicles (PMV), this segmented conception of the city has to change or, at least, reconsider the cut-off thresholds and account for other types of mobility. It is also necessary to mention that when selecting the population that are served by a park, the Euclidean distance, i.e., the distance in a straight line from the nearest park to the place of residence, was used, instead of the real distance using the city's road section. With this latter consideration, accuracy could be increased. On the other hand, the central location of the urban parks with respect to the city as a whole can influence the value of the homes located in more central areas, as opposed to others that are further away. In the case of the quality of the Tarragona parks, we did not find this association, and so the statistical correlation does not follow this trend.

The potential user public in each park (served population) is not made up of a uniform group of individuals with common demographic, economic and social characteristics. In fact, there is a consensus that social injustice regarding urban green areas usually comes along with a certain social stratification and/or residential segregation [47], which can lead to what some authors call "green gentrification" [43]. In this respect, some contributions examine the relationship between the number, proximity and quality of green areas, the socio-economic characteristics of the inhabitants [10,26,44,48] and the composition of the dwellings [48] In addition to these characteristics, consideration is often given to the ethical composition [28,29,49]. For example, De Suosa et al. (2018) [43], in their comparative work between Faro (Portugal) and Tartu (Estonia), observed significant inequalities in the housing districts of the socialist stage in the first of the cities, where most of the Russian minorities live (with a variable availability of public green spaces between 1.04 and 164.07 sqm. per inhabitant), whereas the Romanian communities in Faro were located

in districts without access to public green spaces, although there were smaller differences (from 1.22 to 31.44 sqm. per inhabitant). Other studies, on the other hand, focus on specific demographic groups such as young people [13,50] or immigrants [27]. The availability of this information and its high degree of territorial disaggregation in sources such as the Municipal Register of Inhabitants makes it easy to use, since it can be georeferenced. The data on the population's income are a different case, because, due to confidentiality issues, they are difficult to obtain. In this case, it is essential to use secondary sources to deduce the economic level of the population that a park serves.

The relationship between the disaggregated and precise socio-economic data made it possible to obtain positive correlations between these characteristics, park accessibility and quality, providing evidence of deficits in some areas in the city of Tarragona. Certain urban areas are under-endowed with urban parks vis-à-vis the land occupancy system and are mostly under dispersed forms and without general system reservations, beyond the needs arising from mobility. They require greater research attention. By comparing the different green areas with the neighborhood's social characteristics, it is possible to identify the priority areas and improve their condition, accessibility, quality and distribution [51]. This research has not segmented the socio-economic information on specific groups (according to age, origin), although it has standardized the values used for analyzing them. In future research, it would be relevant to consider the different user groups, the feeling of safety and security and social interaction.

This work has considered the quality of the parks as a factor of environmental justice because urban green areas are relevant for the urban quality of life and for promoting environmental equity [52]. Some authors [53] believe that in order to determine the environmental justice, accessibility and availability of green areas, you have to also consider their quality. The results of their analysis, combined with the socio-economic characteristics, broaden the understanding of environmental justice with respect to the parks. The literature on this issue has found numerous proofs of this. Corley et al. [30] established relationships between the various aspects involved in the quality of urban green areas, and Brown et al. [54] found significant associations between types of urban park and their benefits for the population. Another important conclusion drawn from these approximations is that improving and designing urban parks should consider resident preferences [55,56]. These actions would allow them to become community assets [57].

In response to this need, this work built the PQI: a synthetic index that can be used to assess environmental justice with greater precision than the availability of green surface areas or their distance. Another significant contribution from this work is that the lack of information on the population's income level was substituted by alternative sources. Therefore, to establish correlations with the PQI, the work instead used the level of studies, the resident population's HDI and home sale and rental prices within 300 m of a park. Out of these variables, the one that showed the strongest correlation was home sale prices.

Using the MCE techniques to build the PQI is an attempt to reduce subjectivism, but this always remains because choosing the factors and their weights is a subjective action. Differentiating between the selection of weights (vegetation 40%, facilities 30% and furnishings 30%) can provide a positive reading due to the possible actions aimed at increasing the quality of the parks: improvements to vegetation may require a greater economic effort which, in some cases, has a temporary repercussion (e.g., annual vegetation) or, in others, long-term results were obtained (e.g., time it takes for the tree vegetation to reach adult age). Intervention and maintenance regarding facilities and, particularly, furnishings, can lead to gains in the quality of the parks in a relatively economic way, in a short-term and with significant durability.

#### **6. Conclusions**

There is great disparity among the parks in the city of Tarragona: peripheral parks and central parks; large parks and landscaped squares. In terms of quality, vegetation is the parameter with the best score, while facilities and furnishings have deficiencies.

The value obtained for accessibility is lower than that recommended by the EU, and the same was true for the availability of green spaces, although this reduces if we consider green spaces overall. Based on the correlations between the PQI and the population's socio-economic characteristics, it can be stated that there is a causal relationship between these variables. However, the levels of environmental injustice are reduced. In this sense, in future work, the study could be improved by a more in-depth analysis of the type of vegetation, since certain types of vegetation provide different benefits to the population and, in some cases, can even cause harm; addition of other factors such as landscape quality assessment or new sources as the vegetation index of normalized difference (NDVI), which allows for an estimation of the quantity, quality and development of the vegetation; and, finally, incorporating the user's perception in the PQI model.

This work has shown that the relationship between access to green areas and environmental justice is complex. The aggregate results of the PQI as well as those of the sub-indicators and factors can be used by a competent administration to decide which parks should be a priority and which factors should be improved in each of them. The deficits in accessibility and quality can be overcome with strategies and actions, which will, on the one hand, increase the supply of green spaces in places with inequities and, on the other, increase the quality of these parks, with a greater endowment of cultural, recreational and sport facilities. The new urban agendas, especially in the post-pandemic context, include an analysis of urban habitability, reduction in social inequalities and improvements in health conditions; therefore, the planning, design and management of urban green areas should take socio-spatial attributes into account.

**Author Contributions:** Conceptualization, J.A., J.I.M.M. and Y.P.-A.; methodology, J.A. and Y.P.-A.; software, J.A., J.I.M.M., Y.P.-A. and E.B.P.; validation, J.A. and Y.P.-A.; formal analysis, J.A., J.I.M.M., Y.P.-A. and E.B.P.; investigation, J.A., J.I.M.M., Y.P.-A. and E.B.P.; resources, J.I.M.M.; data curation, J.A., Y.P.-A. and E.B.P.; writing—original draft preparation, J.A., J.I.M.M., Y.P.-A. and E.B.P.; writing review and editing, J.A., J.I.M.M., Y.P.-A. and E.B.P.; visualization, J.A, and E.B.P.; supervision, Y.P.-A.; project administration, Y.P.-A.; funding acquisition, Y.P.-A. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work is part of the research project entitled «El paisaje como valor colectivo. Análisis de su significado, usos y percepción social» (The landscape as a common value. Analysis of its significance, uses and social perception) (CHORA-CSO2017-82411-P), financed by the Ministry of Science, Innovation and Universities (National Programme for Fostering Excellence in Scientific and Technical Research 2018–2020), AEI/FEDER, EU and Department of Research and Universities of the Generalitat of Catalonia (2017SGR22).

**Institutional Review Board Statement:** Ethical review and approval were waived for this study, due to it did not collect critical personal information.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Raw data were generated at Tarragona City Council (Ajuntament de Tarragona). Derived data supporting the findings of this study are available within the article and from the corresponding author J.A. on request.

**Acknowledgments:** We acknowledge Lourdes Llorach de la Peña, Joan Jaume Iniesta Girona (Tarragona City Council/Ajuntament de Tarragona) and Robert Casadevall Camps and Marc Domínguez Mallafré (Universitat Rovira i Virgili) for their co-operation for documenting this work.

**Conflicts of Interest:** The authors declare no conflict of interest.
