Free-roaming cats (Felis silvestris catus) in urban environment (Porto, Portugal)

Índice do artigo

3. Results and Discussion

3.1. Free-roaming cats and their densities

People’s common habit to care for free-roaming cats was witnessed within the city of Porto, as also occurs in many other cities, such as Lisbon metropolitan (Duarte et al., 2010; d’Avila, 2016). In all our study areas, free-roaming cats showed a wide distribution, except in the Porto City Park area where cats were rarely sighted inside the City Park and seems to be concentrated in the outside or limits of the City Park (Figure 3).


art gomes fig3

Figure 3.  Study area with cat, Felis silvestris catus, sightings (location of observations of each cat) represented by dark blue dots. Each map refers to study areas in each year of sampling (Contumil map from 2012, Campo Alegre from 2013, and Porto City Park from 2014). (A) Contumil (limited by blue line); (B) Campo Alegre (limited by yellow line); (C) Porto City Park (limited by green line; City Park limits in orange line); Copyright Google Earth 2017).


Only a small fraction of the cats sighted could be unmistakably identified as domestic cats with owners, by the presence of a collar. Although, in this study, most of the cats sighted are considered feral, these numbers and results should be dealt with careful, because it may include domestic cats with owners, but without collar, which we do not have way to distinguish from the real feral ones. It is also important to note that only in the Porto City Park a considerable number of neutered cats (almost half of the cats sighted; results not shown), recognized by an ear clipping, were seen, contrary to the other two areas.

Our survey revealed high densities and wide distribution of free-roaming cats in Porto and within each study area: 201.14 cats km-2 at Contumil, 130.91 cats km-2 at Campo Alegre, and 33.00 cats km-2 at Porto City Park (Table 3; Figure 3). Porto City Park area having the smallest cat density might be explained by the high number of people walking their pet dogs within the city park limits, thus preventing cat colonization of this area. On the other hand, Contumil is the study area with higher cat density, possibly because it is a residential area located outside the city core and so with less major stressing factors able to affect cat presence. Nevertheless, it is possible that the number of cats (as well as the number of colonies below) in each area is higher and our results are probably still an underestimation of the real cat density, because we are basing our estimates in the cats we saw and counted, on each survey. However, we recognize that we could have not seen some hidden cats due to difficult or impossible access to some areas, or difficult conditions for sightseeing.


Table 3. Cats (number, percentage of cats in groups and cat density) and cat colonies (total number and density) total for each study area and globally considering all total study areas.


Number of cats

Study area (km2)

Cat density

(cats km-2)



% in

Total number










Campo Alegre







Oporto City Park















The number of cats sighted during each sampling period showed some daily variations (Figure 4). However, we did not find any significant differences among different periods of the day (morning vs. afternoon vs. night), neither considering the total number of cats sighted across all the sampling months (Krustal-Wallis test, P = 0.43) nor the cats sighted in each month (Annex Table A1). Therefore, no pattern of cats daily feeding activity was detected, but the periodicity of feeding station refilling (not analysed), might be a factor affecting such activity. Other factors such as temperature and life cycle stage might also be responsible for behavioural changes in cat distribution (not tested). Furthermore, the total number of cats sighted per month showed some variations, as well (Figure 4). February to March are the months with higher cats sighted (Contumil with more cats sighted in February, Campo Alegre in January and February, and Porto City Park in March), which is probably explained by the start of the reproductive period (Mirmovitch, 1995; Natoli and De Vito, 1989; Nowell and Jackson, 1996). We found significant differences among sampling months, either considering each different period of the day (Annex Table A1) or the total number of cats sighted (Krustal-Wallis test, P = 0.004). Thus, this variation in the abundance of cats sighted across different months of the sampling periods are in agreement with previous conclusions that the activity of the domestic cats depends on the season (Gunther and Terkel, 2002).


art gomes fig4

Figure 4. Total number of free-roaming cats sighted during the observation periods along the study. (A) Contumil; (B) Campo Alegre; (C) Porto City Park.


3.2. Free-roaming cats´ colonies distribution

Characteristics of colonies and their habitats, at each study area, are summarized in Table 4 and Annex Table A2. We found that the number of cats in each colony, as well as their occupation range, is highly variable (Table 4). Most of the colonies are small (Campo Alegre: 8 ± 4 cats; and Porto City Park: 4 ± 1 cats), except in Contumil area where most of them are big (16 ± 7 cats). This might be explained, as before, by the location of Contumil area: a suburb area near scrublands and wastelands, where cat numbers have the possibility to increase due to wide areas, while still leaving near people who feed them. With this, it is evident that colony size is significantly smaller in the residential areas located near the city core (Campo Alegre and Porto City Park; Table 5). Therefore, we found that although cats can adapt to a variety of urban environments, they seem to establish their colonies in places near green and open spaces, meaning that besides food and shelter, this is probably an important factor determining colony size.


Table 4. Characteristics of the cat colonies in the study areas (Contumil, Campo Alegre, and Oporto City Park). The location coordinates were obtained from Google Earth and refer to the central point of the colony area (obtained by minimum convex polygon). Cat number is the total number of cats sighted in each colony; Colony area size in ares; colony density within colony area is cats a-1.

Study area  Colony  Location  Colony
Cat number Area (a) Density
Contumil            1 41° 10.228'N / 8° 34.983'W 13 76.83 0.17
2 41° 10.165'N / 8° 34.925'W 18 40.03 0.45
3 41° 9.972'N / 8° 34.717'W 9 94.98 0.09
4 41° 10.475'N / 8° 34.656'W 11 58.37 0.19
5 41° 10.035'N / 8° 35.136'W 23 26.57 0.87
6 41° 10.109'N / 8° 35.162'W 3 7.86 0.38
7 41° 10.739'N / 8° 34.576'W 7 8.20 0.85
8 41° 10.779'N / 8° 34.472'W 15 10.88 1.38
9 41° 10.507'N / 8° 34.229'W 26 13.63 1.91
10 41° 9.971'N / 8° 34.833'W 20 7.78 2.57
11 41° 10.618'N / 8° 34.254'W 26 181.88 0.14
12 41° 10.132'N / 8° 34.975'W 17 22.60 0.75
Campo Alegre 1 41° 9'28.23"N/  8°38'11.98"W 6 23.66 0.25
2 41° 9'7.58"N/  8°38'7.26"W 5 7.69 0.65
3 41° 9'13.54"N/  8°37'58.73"W 7 13.48 0.52
4 41° 9'21.94"N/  8°38'10.74"W 7 72.61 0.10
5 41° 9'24.93"N/  8°38'17.62"W 7 14.06 0.50
6 41° 9'12.73"N/  8°38'2.99"W 4 0.26 15.38
7 41° 9'24.65"N/  8°38'24.14"W 5 3.24 1.54
8 41° 9'32.06"N/  8°38'46.60"W 7 4.64 1.51
9 41° 9'12.93"N/8°37'53.97"W 3 3.93 0.76
Oporto City Park    1 41º 10.234 N / 8º 40.720 W 4 1.90 2.11
2 41º 10.152 N / 8º 41.212 W 5 15.71 0.32
3 41º 10.100 N / 8º 41.171 W 3 3.33 0.90
4 41º 10.073 N / 8º 41.101 W 3 4.26 0.70


The area occupied by colonies at Contumil is 46 ± 52 ares, at Campo Alegre is 14 ± 19 ares, and at Porto City Park is 6 ± 6 ares. It is clear the great variability in these areas estimates (i.e., the high standard deviations of each study area); nevertheless, Contumil (the less urbanized area) and Campo Alegre (the most urbanized area) are again statistically different from each other (Table 5). Colony areas for the remaining study areas comparisons are not significantly different (Table 5). We additionally found that a higher number of cats are significantly found in colonies with higher occupied areas (Spearman’s correlation: β = 0.64; P = <0.001). Finally, cat densities for our study areas are 0.81 ± 0.78 cats a-1 for Contumil, 2.15 ± 4.04 for Campo Alegre, and 1.01 ± 0.77 for Porto City Park. However, these cat densities are not significantly different from each other, among study areas (Table 5).


Table 5. Results of the statistical analyses for differences in the total number of cats sighted in each colony, in the colony area, and in the cats colony density, between the different study areas. Values of the t statistics and p values, from the Students’ t-test are present. See text for mean and standard deviations (SD) information. *Differences statistically significant (p ≤ 0.05)



Differences among study areas (Student’s t-test)


Contumil vs. Campo Alegre areas

Contumil vs. Porto City Park areas

Campo Alegre vs. Porto City Park areas


t (P)

Total number of cats sighted

*5.13 (0.03)

*6.76 (0.02)

2.44 (0.14)

Colony area

*6.76 (0.02)

3.88 (0.07)

0.92 (0.35)

Cat’s density

2.24 (0.15)

0.02 (0.89)

0.76 (0.40)


Furthermore, we showed that cats were mostly seen in groups/colonies in Contumil and Campo Alegre study areas (Table 3). Location of feeding spots, shelter and garbage container, near colonies, is shown in Figure 5. Most feeding spots were clumped (75% in Contumil, 62% in Campo Alegre, and 75% in Oporto City Park; Annex Table A2). Moreover, most of the shelter found for these feral cats is casual rather than specific (92% in Contumil, 85% in Campo Alegre), except for Porto City Park where all the colonies were found with a specific shelter for them (Annex Table A2).


art gomes fig5

Figure 5. Study area with colony locations (central point of the colony area, obtained by minimum convex polygon), known food stations, food type and shelters. Yellow squared marks represent colony location, red circles represent specific shelter, orange circles represent open garbage containers and green circles represent cat lovers’ hand-outs (clumped food distribution). Again, each map refers to study areas in each year of sampling (Contumil map from 2012, Campo Alegre from 2013, and Porto City Park from 2014) (A) Contumil; (B) Campo Alegre; (C) Porto City Park (Copyright Google Earth 2017).


These groups/colonies are clearly associated to the presence of food, either stations regularly filled by cat lovers’ hand-outs (clumped food distribution) or at dumps and refuse depots (dispersed food distribution), because no colony was found far from these food resources (Figure 5; Annex Table A2). However, considering the total number of cats sighted in each colony, its area and densities, we found no differences among different types of food distribution (clumped vs. dispersed; Table 6). Therefore, our study revealed that an important colony factor is the presence of food source, either wittingly or accidentally, which is according to previous studies that also reported that supplemental feeding by people is the main factor for maintain large cat populations in an urban environment (e.g., John, 1977; Izawa et al., 1982; Natoli, 1985; Liberg et al., 2000; Gunther and Terkel, 2002; Flockhart et al., 2016).

Available shelters is not considered a limiting factor, in previous studies, because usually are abundant, although mainly accidental, in urban environments (sensus Bradshaw et al., 1999). However, we found that the number of cats in a colony and the colony area are significantly higher for colonies with accidental shelters than for colonies with specific shelters (Table 6), and one previous study also referred to the availability of shelter as a factor influencing the distribution of cats (Calhoon and Haspel, 1989).

Finally, colony size can also be influenced by urban barriers and traffic kills, as well as by the type of habitat where they stablish. Thus, we further found a significant difference among habitat type, considering the cat colony densities (Table 6).


Table 6. Results of the statistical analyses for differences in the total number of cats sighted in each colony, in the colony area, and in the cats’ colony density, between different categories of food distribution, shelter availability and classes of habitat. For analyses of the differences between food distribution and shelter availability classes, values of the t statistics and p values, from the Students’ t-test, as well as the mean and standard deviations (SD) are presented. For analyses of the differences between habitat type classes, P values from Krustal-Wallis statistical test are presented. † Non-homegeneous variance; * Differences statistically significant (p ≤ 0.05).


Differences between food distribution

Differences between shelter availability

Differences between habitat type


Student’s t-test

Clumped food

Dispersed food

Student’s t-test

Accidental shelter

Specific shelter


t (P)

mean ± SD

t (P)

mean ± SD

Krustal-Wallis Test (P)

Total number of cat sightings

-0.92 (0.37)

9.60 ± 6.95

12.22 ± 7.56

*2.80 (0.01) †

11.91 ± 7.27

5.71 ± 4.19


Colony area

-0.74 (0.48) †

21.44 ± 25.75

36.67 ± 59.27

*2.75 (0.01) †

32.36 ± 42.81

6.72 ± 5.03


Cat’s density

0.53 (0.60)

1.62 ± 3.31

1.02 ± 0.79

0.38 (0.71)

1.55 ± 3.18

1.08 ± 0.62



Considering the total number of colonies in each total area of the study areas, it is clear the difference between the colonies densities of the Porto City Park (3.9 colonies km-2) and the ones at Contumil or Campo Alegre areas (6.8 colonies km-2 and 8.1 colonies km-2, respectively; Table 3). This difference might be explained, as before, by the high number of visitors of the City Park and especially by all the domestic dogs that follow them, preventing the normal spread and establishment of cat populations.


3.3. Urban free-roaming cat estimated densities, in Porto (Portugal), and the "cat problem"

With the surveys of this study, previous number of cats (n=772, 20.42 cats/km2 estimated density) and cat colonies (n=67, 1.8 colonies/km2 estimated number) within the city limits, obtained using local animal welfare NGO’s data (“Animais de Rua” and “Catus Association”), are therefore not surprising (Figure 6). However, such numbers may largely underestimate the real situation. If we were to extrapolate cat densities for the city of Porto, based on the cat densities estimated in this study, we would found far greater values (137.53 cats/km2). However, this estimation may be biased because cats are widely distributed across the city, with different densities, according to food availability, as we shown. Nevertheless, this might be a guideline for future studies, which are needed, especially in different areas or different cities. Considering this extrapolation as a guideline value, our density estimates are thus similar to the ones referred to South Africa, New Zealand and Britain (by Fitzgerald and Veitch, 1985; Jones and Downs, 2011; Sims et al., 2008, respectively; Table 1). Still, some studies on different urban environments found smaller (e.g., Liberg, 1980; Harper, 2004; Kilgour et al., 2017) or much higher (e.g., Izawa et al., 1982; Miramovich, 1991; Flockhart et al., 2016) density estimates, of cat populations (Table 1).


art gomes fig6

Figure 6. Porto city area registering presence of cats and cat colonies, obtained using local animal welfare NGO’s data: Animais de Rua data (2012, unpublished data) in green; Catus association data (2012, unpublished data) in pink (Copyright Google Earth 2017).


Filling feeding spots implies a daily and direct contact between people and free-roaming cats, such contact being potentially harmful to human populations if cats are vectors of diseases. Consequently, cat populations would also be jeopardized due to non-controlled actions against cats, by humans. The “cat problem” is growing and neutering programs are being implemented in the city, both by city municipality and by local animal welfare NGO’s (e.g., “Animais de Rua”; “Catus”), but these might be yet insufficient due to lack of studies revealing the true densities of cat populations. The effort is however higher in the city centre and in places closer to houses, than in the suburbs; the explanation is the higher contact between people and cats, specifically due to the higher number of people feeding free-roaming cats, and to the possibility of cats becoming  pets. This is clear in our results as an higher percentage of neutered cats was found in Porto City Park (41%), which might be due to the proximity to human houses; in fact, although most of the area is green, this study area is surrounded by a dense urban area.

Thus, the implementation of controlled neutering programs, as in other places (e.g., d’Avila, 2016; Kilgour et al., 2017; Tan et al., 2017), would allow to control the spread of diseases to humans and other species (birds, rodents, etc), the impact of feral cats in the predation and extinction of native species, and the probable contact between domestic and wild cats. Rather than just implement these programs in urbanized places near to people, the same arguments also advise the transposition of those measures to other places, such as the city suburbs, forests, and wild areas.