Protected Areas Do Not Outweigh the Species Richness or Functional Diversity and Traits of Birds in Brazil

Simple Summary

Protected areas (PAs) are important for the conservation of fauna and flora. Given that they are established to safeguard these natural resources, it is often assumed that habitats within PAs are better conserved and support a higher diversity of species compared to unprotected habitats. I aimed to test this assumption by studying forest birds in 20 fragments of the Brazilian Seasonal Semideciduous Atlantic Forest, 12 of which are protected by the state. These fragments ranged in size from 43 to 37,000 hectares. I measured species richness and the biological functions provided by these bird species but found no significant differences between protected and unprotected areas. These results may be attributed to the way in which Brazilian protected areas were designated, as they consist of some of the last remaining fragments in a landscape that has experienced severe fragmentation. As a result, in this region, protected areas are not necessarily more intact or better preserved than unprotected forests.

Abstract

Protected areas (PAs) are widely recognized as being essential for maintaining global biodiversity. In this study, I aimed to determine whether PAs support a greater richness of forest bird species than unprotected habitats and whether they exhibit functional diversity indices and traits comparable to those found in protected forests in the eastern Brazilian Semideciduous Atlantic Forest. I conducted bird surveys across 20 forest fragments of varying sizes (43 to 37,000 ha), 12 of which were legally designated as PAs. The results indicated that species richness was not significantly higher in PAs and that there were no significant differences in functional diversity indices and traits between unprotected and protected forests. In eastern Brazil, some forest remnants were designated as PAs merely because they were among the last remaining patches within a heavily fragmented landscape. Therefore, protected areas do not necessarily serve as accurate proxies for well-preserved habitats.

1. Introduction

The establishment of protected areas originally stemmed from a desire to conserve regions with exceptional natural features, such as beauty, grandeur, or rarity [1]. This early conservation focus also included efforts to protect charismatic examples of flora and fauna, such as large trees and esthetically appealing animals [2]. Additionally, there was a growing concern regarding the preservation and the regulation of water sources and their quality for urban populations [1]. The most widely recognized milestone in the development of modern protected area policy was the creation of Yellowstone National Park in the United States in 1872 [2]. Since the late 19th century, national parks have expanded globally, becoming the most established and well-known form of protected natural areas. Over time, the objectives of protected areas have evolved, and they have diversified into various categories that serve distinct purposes which are consistent with the overarching goal of safeguarding nature, which encompasses a range of preservation and conservation objectives, as well as the regulation of resource exploitation [1].

In Brazil, the first recorded proposal for the creation of national parks dates to 1897. However, it was only in 1937 that the country established its first protected area (Itatiaia National Park) [1]. According to the Federal Constitution, protected areas are defined as a territorial space and its environmental resources, including jurisdictional waters, with relevant natural characteristics, legally established by the Public Power, and conservation objectives and defined boundaries under a special administrative regime to which appropriate protection guarantees apply [1]. The National System of Conservation Units (SNUC) Law introduced an innovative approach by classifying protected areas into two broad categories, each with distinct characteristics: integral protection units, which primarily aim to preserve nature while permitting only indirect use of natural resources, and sustainable use units, which seek to balance nature conservation with the sustainable utilization of a portion of its natural resources [1,3]. These management categories align with the criteria set by the International Union for Conservation of Nature (IUCN).

Protected areas (PAs) have been shown to support a substantial proportion of forest-dependent, endemic, and threatened species in both boreal and tropical regions [4,5,6,7]. The critical role of PAs in conservation has sparked discussions about their planning in the context of future climatic changes [8,9]. PAs are also recognized for their importance in metropolitan regions, where habitat modification increasingly converts natural landscapes into conurbations [10]. However, the effectiveness of PA networks must be carefully evaluated. For instance, larger areas are expected to harbor larger bird populations, and PAs should be properly geographically and numerically distributed to include small-ranged species which are common in Neotropical regions [11].

Researchers have investigated how birds’ functional traits vary in response to biotic and abiotic factors, enhancing our understanding and demonstrating that species distribution may not adhere to a consistent pattern [12,13]. In addition, various functional diversity indices have been developed, including functional richness, which represents the total functional space occupied by a community; functional evenness, which describes the uniformity of abundance distribution within the functional trait space; functional divergence, indicating how abundance is distributed along a functional trait axis within the range occupied by the community [14]; and functional dispersion, which measures the average distance of individual species in multidimensional trait space from the community’s centroid [15]. These indices may show random patterns within communities, and, despite relatively stable alpha taxonomy over time, shifts in functional traits can occur [12]. Such community dynamics may be insufficient in regard to maintaining species richness over time, suggesting that additional strategies may be required to preserve ecological functions [16].

The Atlantic Forest is notable for being home to 70% of Brazil’s population, and it is the most extensively studied forest in the country [17]. Nevertheless, significant knowledge gaps persist, particularly concerning tetrapod diversity, with many species still being described, especially in understudied regions and habitats [18]. For example, two closely related bird species were recently discovered only 50–100 km from major urban centers like Curitiba and São Paulo in the states of Paraná and São Paulo, respectively [19,20]. Such findings reinforce the importance of ongoing monitoring, even in well-documented areas [21,22]. The states of Paraná and São Paulo, located in southern Brazil, are largely covered by the Atlantic Forest but are also interspersed with savanna strips, representing the southernmost extent of the Cerrado, another biodiversity hotspot [23,24,25]. The deforestation history of these states, driven primarily by coffee plantation expansion, shows that most inland forests were lost between the 1850s and 1950s [26,27]. Although bird inventories in these states have been conducted for over 200 years [28,29,30], recent surveys have largely overlooked unprotected or previously unsurveyed fragments, as well as revisits to sites that were thoroughly surveyed in the past [28]. Instead, only a few forest fragments have been consistently monitored over time [12,31,32,33,34,35,36].

Climate change, habitat heterogeneity and modification influence species richness and abundance within PAs [9,11,37], which explains why species richness is not always higher inside them [4]. However, in the Atlantic Forest, fine-scale analyses appear to play a crucial role [38,39]. To investigate this, I conducted rapid bird inventories in forest fragments across seven unprotected sites and 10 PAs in Paraná and São Paulo. My main objective was to determine whether PAs support a higher number of forest bird species. I hypothesized that forest bird species richness would differ between PAs and unprotected areas. Additionally, I examined whether unprotected forests exhibit comparable functional diversity and traits to those found inside PAs. I predicted that PAs would harbor more species than unprotected forests [40] and would display higher functional diversity indices and trait values than unprotected forests. This is because I assumed that PAs served as proxies for well-preserved vegetation, which, in turn, shapes the ecological attributes of bird communities [41].

2. Materials and Methods

2.1. Study Areas

The study sites included sixteen municipalities: seven in São Paulo (with seven forest fragments) in southeastern Brazil, and nine in northern Paraná (with 10 forest fragments) in southern Brazil (Figure 1). The inland regions of these states are covered by the most threatened vegetation type of the Atlantic Forest, the Seasonal Semideciduous Forest [24]. These forests, found below 800 m and along the valleys of the Paraná River, comprise riparian, alluvial, and submontane forests [23,24]. The region was settled in the mid-19th century, with severe deforestation beginning as early as the 1850s in São Paulo and expanding in the 1920s in northern Paraná (Norte Pioneiro) to establish large-scale coffee plantations. By the 1920s, most forests had been removed from inland São Paulo [27], and, after 1935, deforestation extended west of the Tibagi River to make way for coffee plantations in Paraná [26]. Currently, 100,882 ha (median: 18.9%; range: 4.3–36.2%) in São Paulo and 74,819 ha (median: 10.8%; range: 5.1–18.7%) in Paraná remain forested in the municipalities surveyed in each state [42]. The region experiences two distinct seasons: a rainy summer (October–March) and a drier winter (April–September), with a humid subtropical climate (Cfa), according to Köppen’s classification, characterized by hot summers [43]. I inventoried 17 forest remnants which varied in size and successional stages (

Table 1. Forest remnants where birds were observed along with their corresponding municipalities, elevation (Alt.) (m), legal protection status, decimal coordinates, area (ha), survey dates, and survey start times. Fragments 1–7 are unprotected areas and fragments 8–17 are protected areas.

ID	Municipality	Alt.	Protected	Latitude	Longitude	Area	Dates	Time
1	Barra do Jacaré	505	No	−23.1	−50.1	43.8	19–20/10/2015	07h40
2	Santo Antônio da Platina	485	No	−23.2	−50.0	90.0	19–21/10/2015	06h00
3	Jacarezinho	410	No	−23.0	−50.0	130.5	19–21/10/2015	07h15
4	Cornélio Procópio	610	No	−23.2	−50.6	299.9	25–27/10/2015	07h30
5	Presidente Epitácio	331	No	−22.2	−52.3	536.0	21/09/2015; 07–08/11/2015	05h40
6	Caiuá	350	No	−21.6	−51.9	1131.0	15–17/09/2015	05h56
7	Anhembi	540	No	−22.7	−48.2	2200.0	14–17/11/2015	05h57
8	Ibiporã	470	Yes	−23.3	−51.0	74.0	28–30/10/2015	09h00
9	Jacarezinho	480	Yes	−23.2	−50.0	163.5	19–21/10/2015	09h00
10	Centenário do Sul	368	Yes	−22.8	−51.5	303.0	31/10; 01–02/11/2015	06h00
11	Santa Rita do Passa Quatro	606	Yes	−21.7	−47.6	353.7	29–30/11/2015; 01/12/2015	05h55
12	Londrina	635	Yes	−23.5	−51.2	675.7	28–30/10/2015	06h00
13	Porto Ferreira	589	Yes	−21.8	−47.4	711.1	29–30/11/2015; 01/12/2015	07h26
14	Santa Mariana	558	Yes	−23.2	−50.6	832.0	25–27/10/2015	06h15
15	Diamante do Norte	315	Yes	−22.6	−52.9	1427.3	03–05/11/2015	06h15
16	Jundiaí	1060	Yes	−23.2	−46.9	2071.0	17–19/11/2015	06h06
17	Teodoro Sampaio	380	Yes	−22.6	−52.3	37,000.0	05–07/11/2015	05h27

, Supplementary Materials, Section S1). All forests are, at least, 35 years old [42].

The objectives and guidelines outlined in the SNUC Law reflect four key concerns: (1) the conservation of biodiversity at three essential levels (genetic diversity, species diversity, and ecosystem diversity), (2) the sustainable use of natural resources, (3) the participation of society in conservation efforts, and (4) the equitable distribution of benefits arising from the creation, implementation, and management of protected areas. All the protected areas I visited were established to meet those broader conservation objectives rather than being designed to benefit a single species, as demonstrated in their management plans [44,45]. Therefore, I treated those protected areas as proxies for preserved habitats.

2.2. Bird Inventories

From 15 September to 1 December 2015, forest fragments were visited primarily over three consecutive days, typically in the mornings (Table 1). I employed 10-species lists, or MacKinnon lists [46,47,48,49], recording the first 10 species observed or heard within a radius of ~50 m, with no species being repeated within the same list. Subsequent lists could include previously detected species but could not repeat any species within the same list. I employed the list method because it has been shown to accumulate species counts more rapidly [50,51] and it is commonly used in the Atlantic Forest for surveying purposes [52,53,54]. The forests I surveyed are depauperate, and remaining stationary during point counts would not have been effective for recording as many species as I achieved while walking along trails. In this method, there are no fixed points; instead, I walked through the environment, compiling a list of 10 species before starting a new list. The total number of accumulated lists served as a measure of sampling effort. Also, there is no fixed number of lists per site. I continued creating lists until no additional species could be detected to complete a 10-species list. This approach typically resulted in surveys lasting around 3 h per visit, amounting to a total of ~153 h. Birds were identified visually using binoculars and by their vocalizations. Unseen individuals whose vocalizations were not immediately identifiable were recorded in the field. I then played back their vocalizations to attract the individuals, enabling subsequent species-level identification. If more than one individual of a species was detected, only a single record was counted.

2.3. Analyses

As I surveyed Atlantic Semideciduous Seasonal Forests, I focused my analyses solely on resident forest-dependent species occurring in the region as classified according to a Neotropical birds reference [55]. Forest species included those classified as F1 (tropical lowland evergreen forests, including edge species—F1E), F4 (montane evergreen forests, including edge species—F4E), F8 (gallery forests, including edge species—F8E), and F15 (secondary forests). I also determined functional traits such as trophic guilds and body masses, sourced from EltonTraits [56], as well as vertical foraging strata and microhabitat specialization [55]. I then estimated functional richness (FRic), functional evenness (FEve), functional divergence (FDiv), and functional dispersion (FDis) using the FD package [14,15], applying regression analyses based on the forest fragment area in ha.

I aimed to assess whether my sampling effort (obtained from rapid assessment data) was sufficient for subsequent analyses, so I generated accumulation curves and judged relevant to calculate a non-parametric estimator (Chao2) using 100 permutations. After verifying the assumptions of the tests, I compared estimated species richness and functional diversity indices between protected and unprotected forests using a Mann–Whitney test and a Multivariate Analysis of Variance, respectively. I further evaluated the effect of elevation and area on the estimated species richness, and whether the conservation status (protected x unprotected) influenced estimated species richness using a Generalized Linear Model (GLM) while considering a Poisson distribution. Thus, I examined whether protection moderates the relationship between elevation area and the estimated species richness. The differences in the proportions of functional traits between protected and unprotected forests were assessed using a χ² test for independence. All analyses were conducted within the R environment [57] (Supplementary Materials, Section S2). The bird taxonomy and nomenclature followed the guidelines of the Brazilian Ornithological Records Committee [58], and the Atlantic Forest endemic species followed a recent proposal [59]. The threatened status at the state level is based on the most recent evaluations for the states of São Paulo [60] and Paraná [61].

3. Results

3.1. Species Richness

I compiled 326 (median: 18.0; range: 10–42) 10-species lists while documenting a total of 165 species (Supplementary Materials, Section S3). Overall, unprotected areas had 135 species, while protected areas had 158. Species accumulation curves indicated slight asymptotic trends, showing no significant relationship with either fragment size or protection status (Figure 2). The median number of species per fragment was 41.0 (29–74), while species estimations exhibited a broader range, with a median of 57.4 (33.2–110.2) (

Table 2. Number of accumulated 10-species lists, number of Atlantic Forest endemic species, empirical (S) and estimated (Chao) forest species richness, and functional diversity indices per fragment. The indices include FRic (functional richness), FEve (functional evenness), FDiv (functional divergence), and FDis (functional dispersion). Fragments 1–7 are unprotected and fragments 8–17 are protected areas.

Fragment	Lists	Endemics	S	Chao	FRic	FEve	FDiv	FDis
1	10	3	30	33.2	0.112	0.785	0.712	0.180
2	18	6	41	44.8	0.136	0.769	0.844	0.224
3	18	6	40	41.5	0.194	0.781	0.761	0.205
4	21	12	53	58.8	0.287	0.809	0.850	0.196
5	15	2	37	46.1	0.136	0.745	0.746	0.214
6	10	0	30	36.1	0.119	0.760	0.702	0.203
7	42	19	71	82.4	0.262	0.641	0.752	0.191
8	19	7	38	44.4	0.115	0.744	0.684	0.164
9	17	5	40	48.5	0.069	0.731	0.820	0.154
10	10	1	29	36.5	0.080	0.757	0.795	0.227
11	27	11	61	64.4	0.225	0.752	0.773	0.228
12	33	19	74	100.2	0.296	0.753	0.864	0.193
13	17	13	58	74.3	0.258	0.804	0.725	0.197
14	31	12	66	87.3	0.221	0.693	0.879	0.202
15	11	3	41	57.4	0.097	0.778	0.720	0.190
16	36	18	71	110.2	0.271	0.665	0.883	0.224
17	23	8	55	63.7	0.258	0.750	0.699	0.222

). These estimations represented a median of 83.2% of the observed species richness, varying between 64% and 96%. One species was found in all fragments (Picazuro Pigeon Patagioenas picazuro), 44 species (27%) were observed in only one fragment, while 97 species (59%) were recorded in five or fewer fragments. Overall, I identified 39 Atlantic Forest endemic species, with a median of 7.0 (0–19) per fragment. Additionally, none of the species were listed as globally threatened (IUCN), so I used local state red lists. I recorded four threatened species across four fragments in São Paulo and two threatened species in three fragments in Paraná. Collectively, these six threatened species accounted for 11 contacts (<1%) in the lists, and I will not discuss them further in this analysis.

While there is a tendency for protected areas to support a higher number of forest species, the difference in estimated species richness between protected and unprotected areas is not statistically significant (W = 16.0, p = 0.070) (Figure 3).

Regarding the GLM, the median residual was low (0.21), suggesting a reasonably balanced fit to the model, and the estimate coefficient was highly significant (2.95, p < 0.001). Elevation (0.001, p = 0.005) and area (0.0002, p < 0.001) had a significant, although slight, positive effect on species richness. Protection alone did not show a statistically significant effect on species richness (0.573, p = 0.057). The interaction between elevation and protection was not significant (−0.0004, p = 0.494), meaning that the effect of altitude on species richness is similar for both protected and unprotected fragments. Area, however, had a weak positive effect on species richness for protected areas compared to unprotected ones (−0.0002, p < 0.001), as indicated by the negative interaction term.

3.2. Functional Diversity

Functional richness and dispersion exhibited a slight increase with increasing fragment area, while functional evenness and divergence displayed the opposite trend (Table 2, Figure 4). However, there were no significant differences in any of the functional diversity indices between protected and unprotected forests (F = 2.2, df = 1, p = 0.119). Additionally, there was no clear evidence that protected areas constitute a cluster of legally protected environments (Figure 5).

3.3. Functional Traits

Qualitatively, all functional traits were represented heterogeneously across the 17 fragments (Supplementary Materials, Section S4, Figure S1). The distribution of trophic categories did not differ significantly between unprotected and protected areas (χ² = 2.8, df = 4, p = 0.575), use of vertical stratification (χ² = 0.1, df = 3, p = 0.982), or microhabitat speciality (χ² = 1.5, df = 5, p = 0.911). Unprotected areas exhibited a similar median body mass of 32.4 g (range: 4.1–3400 g) compared to protected areas, which had a median body mass of 28.4 g (4.1–1386.4 g). This difference was not significant (W = 82192.0, p = 0.610). Furthermore, the multivariate analysis of all functional traits revealed no significant differences between protection status (F = 1.2, df = 4, p = 0.372).

4. Discussion

4.1. Species Richness

Given the accuracy of the Chao estimation, I deemed my rapid inventories sufficient to represent most forest species likely to be found across all forest fragments. Protected areas did not harbor a greater estimated number of species than unprotected forests, as has been demonstrated on a broader scale [40], even when controlling for area and elevation. This finding may be attributed to the varying histories of vegetation succession among the fragments. Although some fragments lack documented histories, I observed that smaller remnants tend to be more significantly modified. Consequently, anthropogenic activities play a crucial role in shaping taxonomic diversity within forest fragments [41]. In addition, unprotected fragments support the impact of forest reduction on forest species richness [62]. Further investigations could examine potential reasons why protection and elevation do not interact significantly. The positive influence of area on species richness in unprotected fragments is well documented [63], but the attenuated effect of area in protected fragments warrants further investigation. By considering the land-use history of protected areas in São Paulo and Paraná, I propose that these areas may represent regenerated habitats that no longer support the forest bird communities that once inhabited them. This pattern has been observed in the inland Seasonal Atlantic Forests of São Paulo, where similar trends have been documented [64].

4.2. Functional Diversity

The increase in FRic with larger fragment areas appears to be a common pattern observed both across entire communities [63] and within specific trophic guilds, such as frugivores [65]. This index has also been associated with the presence of bird species that have limited dispersal abilities and are less likely to occupy diverse habitats [66]. I observed that most sites with higher species richness also exhibited higher FRic values, a trend that has been previously documented for forest birds in Paraná [66]. In deforested landscapes, generalist species may supplant forest species, resulting in lower functional redundancy [67]. However, the lower FRic values recorded in some protected areas may be attributed to the structural simplicity and homogeneity of their vegetation, which is often less complex than that found in simplified habitats [68]. The functional diversity and redundancy of bird species are not significantly impacted by forest loss in the Atlantic Forest [69]. Additionally, the high and comparable FRic indices observed in unprotected fragments suggest that these areas may reflect similar niche breadths as those found in protected areas, indicating that species may exhibit reduced capacities in regard to responding to environmental changes [70].

Functional evenness has been shown to have no effect in relation to fragment area [63]; however, I observed that it decreases with increasing fragment size, with no association to the protection status of the forest fragments. Thus, despite the area of the fragments, my findings indicate a reduction in species functions irrespective of legal protection. Notably, FEve may increase with decreasing fragment sizes, which suggests negative implications for functional redundancy and ecosystem resilience [71] due to a potential lack of functionally rare species. Even in fragments as large as 250 ha, I found that FEve exhibited no association with fragment area, raising concerns about the risk of local extinctions [63]. Low FDiv values reflect the prevalence of common species [63], a pattern I found to be unrelated to fragment area. Conversely, higher FDiv values strongly suggest the presence of underutilized forests, indicating that the available ecosystem resources are being used more effectively, meaning that the functions each species provides are distinct from one another [72]. I found lower FDis values in smaller fragments, regardless of their protection status, suggesting that species functions are more widely distributed within these communities, akin to the low functional redundancy observed in deforested landscapes [67]. This finding indicates that the species persisting in smaller fragments exhibit greater ecological differentiation from one another than would be expected by chance [13,73].

4.3. Functional Traits

Not all PAs exhibited a significant representativeness of functional traits, and some were comparable to unprotected fragments. Additionally, functional groups with lower species richness are considered more vulnerable, and my results indicate their absence in certain fragments, likely as a consequence of landscape modifications [74]. This suggests a shift in taxonomic or functional properties [70], indicative of fragments that have been impacted by deforestation, as evidenced by changes in the functions provided by these ecosystems [75]. When analyzing each forest fragment individually, it is important to note that a low representation of understory species may signal the absence of that forest stratum, leading to a lack of correlation between foraging strata and vegetation variables [41]. Furthermore, the lower number of forest species within functional traits may suggest that these communities exhibit resilience following ecological filtering processes caused by habitat modifications, resulting in communities dominated by generalist species [75]. Importantly, none of these conclusions could be attributed to the protected or unprotected status of the forest fragments.

Higher body masses are negatively impacted by a reduction in fragment area, as larger species tend to decline or disappear from smaller fragments [65]. The lack of an area effect on biomass suggests that these forests may harbor fewer functionally redundant species composed of resilient individuals [74]. However, the presence of both large- and small-bodied species appears to be compensated, at least among forest and non-forest frugivorous birds [75]. Consequently, I propose that the absence of a correlation between greater body mass and the area of a PA may result from habitat modification and/or hunting pressure [76].

Researchers have demonstrated the critical importance of large PAs for species conservation through comparisons with unprotected areas [5,6]. Additionally, the negative impacts associated with small PAs often outweigh the benefits of protection, primarily due to limited competition for resources and breeding habitats, increased edge exposure, and various anthropogenic threats [40]. To address these shortcomings, it is essential to sustain the ecological processes facilitated by birds by maintaining high levels of forest cover and minimizing fragmentation [67]. The Atlantic Forest in São Paulo and Paraná is highly fragmented, and conservation efforts must be framed within this context.

Although legally protected, the effectiveness of such protections may be limited in practice due to inadequate law enforcement in certain regions. In contrast, privately owned areas of native vegetation may benefit from more effective conservation practices. For example, unprotected fragments within actively managed landscapes have been shown to play a crucial role in maintaining viable populations of the palm Euterpe edulis when compared to a designated Atlantic Forest protected area [77]. In addition, fragments often resemble the structure and composition of early-successional forests, driving spatial turnover patterns that favor disturbance-adapted species at the expense of primary-forest specialists [78]. However, small fragments can still harbor significant biodiversity and ecosystem functions, making their retention a vital conservation strategy within the highly threatened Brazilian Atlantic Forest [13,72]. For birds, there is a clear consensus that smaller forest fragments should be preserved, as they positively contribute to the maintenance of forest species, provided they are designated as protected areas [40]. In some instances, however, certain PAs may have been established without proper evaluations of their biodiversity, as the previous severe fragmentation likely led authorities to designate only a few remaining forest fragments as protected. Consequently, unprotected forests appear to merit legal protection to prevent the compromise of essential ecological processes.

4.4. Future Directions

The continued monitoring of these 17 fragments is highly recommended. It might be useful to include additional predictors of species richness in future investigations, such as vegetation structure and matrix habitats. The long-term surveys may detect species of conservations concern, such as endemic and threatened species. Since bird diversity is influenced by the size of the island (smaller remnants) and its distance from the mainland (larger remnants) within a fragmented area, it would be valuable to incorporate landscape metrics to assess how matrix complexity impacts the forest birds in this region. The forest-edge ecotone effect, which enhances the diversity of available ecological niches and attracts specific forest-edge species, would also be a valuable consideration for future studies. Finally, further research should address the issue of imperfect detection and consider applying occupancy models to better understand bird occupancy in this Atlantic Forest region.

5. Conclusions

In the states of São Paulo and Paraná, PAs do not always effectively maintain forest bird species richness and functional diversity at levels comparable to those found in conservation units, as some unprotected fragments surpass these metrics. Future considerations for the designation of PAs should prioritize the quality of vegetation rather than solely focusing on fragment areas. Failing to do so could lead to the establishment of PAs in fragments that do not support essential ecological processes, undermining the fundamental objective of legally protected habitats.