https://orcid.org/0000-0002-5288-4106
Figures
Abstract
Large forested tracts are increasingly rare in the tropics, where conservation managers are often presented with the challenge of preserving biodiversity in small and isolated fragments. The Atlantic Forest is one of the world’s most important biodiversity hotspots, jeopardized by habitat loss and fragmentation. The Pernambuco Endemism Center (PEC) is the most degraded of the Atlantic Forest regions and because of the dramatic levels of deforestation, fragmentation, and ongoing species losses, studies on the distribution and configuration of the PEC’s forest cover are necessary. However, across dynamic tropical landscapes, investigating changes over time is essential because it may reveal trends in forest quality attributes. Here, we used Google Earth Engine to assess land use and land cover data from MapBiomas ranging from 1985 to 2020 to calculate current landscape metrics and to reveal for the first time the spatiotemporal dynamics of the PEC’s forests. We identified a forest cover area that ranged from 571,661 ha in 1985 to 539,877 ha in 2020, and about 90% of the fragments were smaller than 10 ha. The average fragment size was about 11 ha, and only four fragments had more than 5,000 ha. Deforestation was mostly concentrated in northern Alagoas, southern Pernambuco, and non-coastal Paraíba and Rio Grande do Norte. On average, borders represented 53.6% of the forests from 1985 to 2020, and younger forests covered 52.3% of the area in 2017, revealing a vegetation rejuvenation process 2.5 times higher than in total Atlantic Forest. In 2017, older forest cores in fragments larger than 1000 ha (i.e., higher-quality habitats) represented only 12% of the remaining forests. We recommend that the amount of forest cover alone may poorly assist conservation managers, and our results indicate that ensuring legal protection and increasing surveillance of the PEC’s few last higher-quality habitats is urgently needed.
Citation: Dias TdC, Silveira LF, Francisco MR (2023) Spatiotemporal dynamics reveals forest rejuvenation, fragmentation, and edge effects in an Atlantic Forest hotspot, the Pernambuco Endemism Center, northeastern Brazil. PLoS ONE 18(9): e0291234. https://doi.org/10.1371/journal.pone.0291234
Editor: Bijeesh Kozhikkodan Veettil, Van Lang University: Truong Dai hoc Van Lang, VIET NAM
Received: April 24, 2023; Accepted: August 24, 2023; Published: September 8, 2023
Copyright: © 2023 Dias et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This study is part of the ARCA Project and received financial support from the Fundação de Amparo à Pesquisa do Estado de São Paulo – FAPESP ARCA project (www.fapesp.br, grant number 2017/23548–2). TCD received a fellowship from Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPQ (www.gov.br/cnpq/pt-br, Procs# 141084/2019–7); MRF and LFS received productivity research fellowships from the Conselho Nacional de Desenvolvimento Científico e Tecnológico – CNPQ (www.gov.br/cnpq/pt-br, Procs# 308702/2019–0 and 308337/2019–0). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Tropical forests harbor more than half of all known species and are among the most endangered ecosystems on Earth [1, 2], with deforestation rates reaching around 5.5 Mha/yr-1 [3]. Besides reducing critical habitats for many species, deforestation also increases fragmentation and edge effects through the conversion of larger blocks of forests into smaller ones [4, 5]. Because habitat loss and fragmentation are the leading causes of wildlife erosion and ecosystem functioning depletion worldwide [6, 7], there is a broad consensus that protecting large areas of natural vegetation is the best way to achieve conservation purposes [8, 9]. However, large forested tracts have become increasingly rare in tropical forest hotspots, where conservation managers are often presented with the challenge of preserving biodiversity in small and isolated fragments [8, 10, 11].
Assessing the spatial patterns and distribution of the fragments in highly degraded habitats is important for conservation planning because they can predict the extinction risks of biodiversity components [12, 13], and because they can indicate the best remnants for the creation of protected areas [14, 15]. However, in dynamic human-modified landscapes, the replacement of older forests by secondary and younger habitats has been recently proved to be a secretive temporal effect that can masquerade the loss of high-quality habitats [16]. In these circumstances, simple assessments of land cover patterns may provide only part of the information needed for conservation planning due to the lack of the temporal component [17]. Investigating landscape changes over time is essential because it can reveal trends in deforestation and forest regeneration rates, and because temporal changes in fragment attributes (e.g., isolation, area, age, and edge effects) influence biodiversity distribution [18].
The Brazilian Atlantic Forest is one of the world’s most important biodiversity hotspots and it has been through a history of intense degradation since the European settlement more than 500 years ago [19–21]. Currently, only about 11.26% of its original 150 Mha remains, almost entirely (80%) in fragments smaller than 50 ha [22]. Despite the continuous deforestation process, it was recently revealed that the amount of vegetation cover was relatively constant during the last 30 years (around 28 Mha) due to a hidden process of substitution of old forests by areas of secondary vegetation, with losses of older habitats ranging from 220,000 to 80,000 ha/year-1 from 2000 to 2015 [16].
The Pernambuco Endemism Center (hereafter PEC) formerly comprised a 4.4 Mha area located north of the São Francisco River [23, 24], in northeastern Brazil [22]. Today, the PEC is the most degraded of the Atlantic Forest regions [22], challenging conservation managers due to the high concentration of threatened endemic taxa [25–28], the reason why it has been considered a hotspot within a hotspot [29]. About 360,455 ha of native forests were detected in the PEC using data from 2005, with about 60% of the total cover influenced by edge effects and the absence of fragments larger than 10,000 ha [22]. Another estimate carried out with data from 2001 to 2007 revealed 322,372 ha of forests, distributed into fragments smaller than 10 ha (38,504 ha), larger than 10 ha (121,081 ha), larger than 100 ha (114,440 ha), and larger than 1,000 ha (48,347 ha) [29].
Despite the efforts to characterize the distribution of the PEC’s habitats during the last decades [19, 22, 29, 30], previous studies did not capture the temporal dynamics of its forest cover. The variation in methodological procedures, associated with the fact that the last comprehensive study was conducted using data from 2007, has impeded proper interpretations about the temporal changes that occurred in the PEC’s forest cover, and there is a lack of information on the recent status of its habitats. Additionally, regional assessments on parameters related to habitat quality (e.g., vegetation age) over the Atlantic Forest are urgently needed for landscape management [16].
Here, we aimed to reveal the spatiotemporal dynamics of the most endangered habitat in the Americas during the last decades, providing an updated descriptive assessment of the changes that occurred in the configuration and distribution of the PEC’s habitats, as well as inferring the current state of its forests. We used Google Earth Engine (GEE) [31] to process MapBiomas annual land use and land cover layers [32] from 1985 to 2020 [32] and we calculated metrics related to the (i) amount of forest cover, (ii) number of fragments, (iii) average fragment size, (iv) largest fragments, (v) deforestation rates, (vi) forest regeneration rates, (vii) amount of core and edge areas, and the (vii) amount of older and younger forests. We predicted that the forest rejuvenation dynamic reported for the whole Atlantic Forest [16] could be even more remarkable in the PEC because differently from southern Atlantic Forest regions, large and well-preserved forest tracts no longer exist northern from the São Francisco River [22]. In this study, we incorporate forest quality parameters not previously considered by conservation managers and we developed the open-access GEE application “Forests of the PEC” (https://dias93thiago.users.earthengine.app/view/forests-of-the-pec) that may potentially assist conservation planning in this important biodiversity hotspot.
Methods
Description of the study area
The PEC is a biogeographic zone of the Brazilian Atlantic Forest located in northeastern Brazil across the states of Alagoas (AL), Pernambuco (PE), Paraíba (PB), and Rio Grande do Norte (RN) (Fig 1). Human interferences were responsible for the intense degradation of its forest formations since the 16th century, and the highest levels of fragmentation and deforestation occurred during the 1970s, mainly due to the activities of the sugar cane industry [33]. This intensive degradation reduced the forest cover to no more than 11.5% of its original size, mostly unprotected (protected areas account for about 1% of the forest cover) [22]. The PEC is located in a portion of the tropical zone where the predominant climate is Köppen’s As (tropical with a dry season), and smaller portions of the climate are Köppen’s Af (tropical without a dry season) are also found. Annual rainfall ranges from 1900–2200 mm in the coastal regions and 700–1000 mm in the western border, with annual mean temperature varying from 24 to 26° C [34]. The PEC went through four major waves of deforestation, the first three occurred between the sixteenth and eighteenth centuries, and the last started after 1975, during the establishment of the Proálcool Program (Brazilian Alcohol Program) [23]. The PEC is home to the highest number of globally threatened species in the Americas [26, 28, 35, 36]. Large mammals are currently extinct in the PEC (e.g., Puma Puma concolor, Jaguar Panthera onca, White-lipped Peccary Tayassu pecari, Brazilian Tapir Tapirus terrestris, Giant Ant-eater Myrmecophaga tridactyla, Gray Brocket Mazama gouazoubira), and at least half of its medium-sized mammals have disappeared over the last 500 years [29, 37, 38]. It was also at the PEC that the modern Brazilian bird extinctions were registered [26, 39, 40].
The 2020 land use and land cover layer from MapBiomas Project [32] was adapted to show relevant aggregations of land cover classes. The class “forest formation” from annual layers represents a proxy of the PEC forest cover. The table on the right shows the identification (ID) and name of its municipalities. The additional layer representing administrative boundaries used in this map was obtained from the freely-available dataset provided by the Brazilian Institute of Geography and Statistics (IBGE).
Dataset
We used GEE to assess data from the sixth collection of the open-source MapBiomas Project, a collaborative initiative that provides annual (1985–2020) land use and land cover information at 30 m spatial resolution for Brazil using Landsat imagery [32]. In GEE, we retained only areas classified as forest formations, representing proxies of the Brazilian Atlantic Forest cover of the PEC. In the MapBiomas Project, all data was processed with methods that accounted for interferences from clouds or atmospheric haze, and several spatial and temporal post-classification filters were applied to ensure data quality [32]. According to MapBiomas, the Collection 6 Level 2 land use and land cover classes of the Atlantic Forest biome has a global accuracy of 85.5%, allocation disagreement of 8.3%, and area disagreement of 6.2% [32]. For the ‘forest formation’ class in Atlantic Forest biome (used in this study), overall accuracy ranged from 85.6% to 87.89% between 1985 e 2018 [32]. A stratified sample design that took the probabilities of sample weight adjustment into account was used by the MapBiomas team (three experts) to validate data using more than 12,000 points [32]. Following the minimum forest size (0.5 ha) established by FAO’s Global Forest Resource Assessment [41], we removed all fragments with less than six pixels (0.54 ha) using Rook’s case for pixel adjacency [42]. Using the ‘clumpSize’ function in ‘bfastSpatial’ package [43] in R [44], we detected clumps of connected cells, assigned unique identification numbers (IDs) for each clump (forest fragment), and calculate their sizes based on the number of connected pixels, which were lastly converted to hectares [43]. For temporal analyses, we selected a time threshold of 35 years (1985–2020) due to the availability of the annual data from MapBiomas, which is the open-access data source with the highest temporal resolution for Brazil [43]. All layers resulting from this study are freely available in the GEE application “Forests of the PEC”, and can be downloaded by assessing the following link https://dias93thiago.users.earthengine.app/view/forests-of-the-pec. Additionally, all of the GEE and R scripts were provided in Supplementary Material.
Forest cover, number of fragments, average, and largest fragment sizes
In GEE, we evaluated temporal changes in the forest cover extent and composition by calculating the overall forest area and the areas of fragments > 10 ha (very small), 10–100 ha (small), 100–1,000 ha (medium), and > 1,000 ha (large). For comparative purposes, we followed the fragment size classes used in previous works on the PEC [29]. We also calculated the total number of fragments, and the number of fragments according to the abovementioned size classes. For forest area and number of fragments, we identified linear trends over time by applying a bottom-up breakpoint analysis using segmented package [45] in R. After selecting the number and location of breakpoints using Bayesian Information Criterion (BIC), this analysis subdivides a time series into phases with distinct trends and slopes [46]. We determined the average fragment size by dividing the total forest cover area by the number of fragments, and we inferred the size of PEC’s largest fragment. To assess the current distribution of forests for each municipality of the PEC in 2020, we generated layers representing the forest cover area (total) and the number of fragments (very small, small, medium, and large fragments). We chose to calculate metrics for each municipality to facilitate the interpretation of our results by state government environmental agencies and policy makers. To capture processes that may operate in different spatial scales, we developed a tool in the GEE application “Forests of the PEC” which permits regions of interest to be defined by users. We additionally identified the remaining largest fragments of the PEC (fragments larger than 5,000 ha) in 2020 and we generated information on their core and edge areas, as well as the areas composed by older and younger forests (see sections below).
Deforestation, forest regeneration, and identification of older and younger forests
We assessed the current amount and the distribution of older (> 35 years) and younger (< 35 years) forests of the PEC in 2017 using GEE. We used the pixels classified as forests in 1985 with no event of deforestation registered until 2017 as proxies of the older forests of the PEC (see methodology for deforestation classification below). Pixels classified as forests in 2017 that did not match these conditions were classified as younger forests. We generated layers containing the overall current distribution of older and younger forests, and their current distributions per municipality. To measure deforestation and forest regeneration from 1987 to 2017, we implemented a moving window-based temporal filter in GEE [16, 47].
Year-specific deforestation events were assumed when a given pixel was classified as forest for the two previous years (t– 2, t– 1) and as non-forest in the current (t) and subsequent year (t + 1) (S1 Fig) [16]. We classified year-specific forest regeneration events when pixels were classified as non-forest for the two previous years (t– 2, t– 1), and then as forest in the current year (t) and in the next three subsequent years (t + 1, t + 2, t + 3) (S2 Fig) [16]. Moreover, we investigated deforestation events across a range of fragment sizes (S3 Fig). Using the annual deforestation layers, we extracted and averaged the fragment size during the two previous years (t– 2, t– 1) for each deforestation pixel. We then classified the resulting pixels of deforestation into deforestation of very small, small, medium, and large fragments.
We discriminated deforestation of older forests by calculating the year of the first deforestation event in pixels classified as forests in 1985. We only classified the first deforestation event of a given 1985 forest pixel as deforestation of older forests. For each year, we remapped the values of deforestation from one to the value corresponding to its year of occurrence. After that, for each pixel, we extracted the minimum value (corresponding to the first event of deforestation) and masked the new classified images to remove all pixels not classified as forests in 1985 (S4 Fig). We classified all the other deforestation events that did not matched the abovementioned conditions as deforestation of younger forests. We addressed deforestation rates of older forests only after 2000 because at least part of the deforestation that occurred in the first years of our study time may stand as deforestation of younger forests that grew just before 1985 [16]. To evaluate the participation of each municipality in forest loss and gain rates across the PEC, we additionally mapped their accumulations from 1987 to 2017.
Core and edge areas
We calculated the Euclidean distance between the forest edge and its interior in GEE to assess changes in the amount of forest cores and edges for the PEC forests. We used a conservative threshold of 50 m to define forest edges since this is the distance in which the vegetation dynamics and structure are more affected by edge effects in the PEC [19, 48–50]. Using the core and edge layers, we calculated the total core and total edge area, as well as the proportions of forests covered by edges and core areas per year.
Results
Forest cover, number of fragments, average, and largest fragment sizes
During the last decades, nearly 5% of the PEC forest cover was lost (571,661 ha in 1985 and 539,877 ha in 2020), and the most drastic reduction occurred before the 1990s (Fig 2A). The amount of forests we detected in 2020 represents 12.3% of the original PEC’s forest cover. Very small fragments (< 10 ha) represented around 25% (144,108 ha) of the forest cover in 1985 and 17% (94,091 ha) in 2020 (S1 Table). For forest areas, the first breakpoint was detected between 1988 and 1991 for total and all other classes of fragment size (Fig 2A), where higher rates of decrease were identified (S3 Table). Decreases in the forest extent were generally found to occur at a lower rate during the 1990s, followed by increases in the forest amount after the 2000s (S3 Table).
Section (A) displays changes in the total forest cover area (total and according to fragment size). Section (B) shows changes in the number of fragments (very small, small, medium, large, and total) from 1985 to 2020. Dotted horizontal lines represent the historical means and dotted vertical lines represent breakpoints. Bold black lines represent linear trends.
Almost all of the PEC fragments (roughly 87.44%) were classified as very small on average during the last decades, and only about 0.08% of the fragments were larger than 1,000 ha. Decreases in the number of fragments were found to be more pronounced during the 1980s, and the PEC experienced a period of lower decreasing rates during 1990 (S3 Table). After the 2000s, increases in the number of fragments were found for all fragment classes and total, with generally high rates of increase for medium and large fragments (S3 Table).
Only five municipalities of the PEC currently maintained more than 10,000 ha of native forests in 2020: Coruripe/AL, Maceió/AL, Murici/AL, Igarassu/PE, and Santa Rita/PB (Fig 3A). Fragments with increased sizes were generally confined to municipalities closer to the coast (eastern PEC) (Fig 3B). Furthermore, only the municipalities of Maceió and Coruripe in the state of Alagoas harbored more than five fragments larger than 1,000 ha each in 2020 (Fig 3B).
Current status of the (A) total forest cover and (B) the numbers of fragments according to their sizes: very small (< 10 ha), small (10–100 ha), medium (100–1,000 ha), and large (> 1,000 ha), for each municipality of the Pernambuco Endemism Center. Fragments located in borders between municipalities were accounted for each of them. The identification of the municipalities is provided in the Fig 1. The layer of administrative boundaries used in this map was obtained from the open-access dataset provided by the Brazilian Institute of Geography and Statistics (IBGE).
The average size of the PEC’s fragments was roughly 10.37 ha over the last decades and a significant increase was observed over time (0.11 ha/yr-1, R2 = 0.68; p-value < 0.001). The average size of the PEC’s largest fragment over time was around 11,812 ha, and in 2020 the largest fragment was bigger than 14,000 ha for the first time since 1986. We identified four forest fragments larger than 5,000 ha in 2017, located in the states of Alagoas (ID 79332), Pernambuco (IDs 38316, 40142), and Paraíba (ID 17378) (Fig 4). The largest fragment identified in 2017 was ID 388316, located in Pernambuco state, with 12,897 ha (Fig 4). All fragments except ID 40142 were primarily composed by older forests that grew before 1985 (ID 79332: 71%; ID 40142: 49%; ID 38316: 72%; ID 17378: 81%) (Fig 4). In addition, ID 40142 also presented the lowest percentage of core area cover (63%). However, except for ID 79332, all fragments larger than 5,000 ha were located nearby roads and state capitals (Fig 4).
Identification of the fragments larger than 5,000 ha in the year of 2017 over Pernambuco Endemism Center. The administrative boundaries, land cover, and federal road layers used in this map were obtained from the open-access datasets provided by the Brazilian Institute of Geography and Statistics (IBGE), MapBiomas Project, and Brazilian Water Agency (ANA), respectively.
Identification of older and younger forests
We identified three municipalities in Alagoas state (Coruripe, Maceió, and Murici), three in Pernambuco (Água Preta, Abreu e Lima, and Igarassu), and two in Paraíba (Santa Rita and Rio Tinto) with more than 5,000 ha of older forests. In summary, we detected a current cover of 237,708 ha of older forests and 260,984 ha of younger forests in 2017 (S5 Fig).
Deforestation and forest regeneration
We identified the accumulated loss of more than 670,000 ha of forests in the PEC over the last decades (Fig 5A). The average annual deforestation rate was around 21,648 ha/yr-1, and the highest losses were observed before the 1990s (1987: 95,198 ha; 1988: 55,486 ha; 1989: 51,595 ha) (Fig 5A). After the 2000s, deforestation rates were mainly below the average annual deforestation rate (Fig 5A). The deforestation rate of older forests was around 1,903 ha/yr-1 since 2000 and of younger forests was roughly 11,159 ha/yr-1 since 1987 (Fig 5A). Deforestation was more common in very small fragments; however, we detected evidence of losses of 20,138 ha in large fragments from 1987 to 1989 (Fig 5B).
Section (A) shows an overview of the annual deforestation rates (total, older, and younger forest losses) and the deforestation accumulation from 1987 to 2017. Section (B) displays the losses in very small (< 10 ha), small (10–100 ha), medium (100–1,000 ha), and large (> 1,000 ha) fragments. The dotted horizontal line is the average annual deforestation rate. Section (C) shows an overview of the deforestation accumulation for each municipality of the PEC, from 1987 to 2017. The municipalities’ identification is provided in Fig 1. The layer of administrative boundaries used in this map was obtained from the open-access dataset provided by the Brazilian Institute of Geography and Statistics (IBGE).
Five municipalities (Penedo/AL, Bonito/PE, Sapé/PB, Pedro Velho/RN, and São José de Mipibu/RN) lost over 10,000 ha of forests since the late 1980s (Fig 5C). Deforestation was largely concentrated in the municipalities near the borders between northern Alagoas and southern Pernambuco, and in a latitudinal gradient extending from the non-coastal PEC regions of Paraíba to middle Rio Grande do Norte (Fig 5C).
Over the last few decades, the PEC experienced the accumulation of 443,324 ha of forest regeneration, with an average annual rate of 14,301 ha/yr-1 (Fig 6A). In general, higher reforestation rates were detected during the first years of our time-lapse (Fig 6A). The municipalities of Maceió/AL, Bonito/PE, and Cabo de Santo Agostinho/PE accumulated the highest forest regeneration rates since 1987 (Fig 6B).
Section (A) displays annual forest regeneration rates and the accumulation of forest gain over time. The dotted horizontal line represents the average annual forest regeneration rate. Section (B) shows the spatial distribution of forest regeneration accumulation across the municipalities of the PEC. The municipalities’ identification is provided in Fig 1. The layer of administrative boundaries used in this map was obtained from the open-access dataset provided by the Brazilian Institute of Geography and Statistics (IBGE).
Core and edge areas
On average, more than half of the PEC forest cover (53.6% ± 3.2%) was located within the first 50 m from the edge, with mean core and edge areas being around 211,615 ha and 245,314 ha, respectively. We found about 10% of increase in the proportion of core areas over our study period. Currently, the total forest cover represented by core and edge areas are 255,228 ha and 284,649 ha, respectively (Fig 7).
Section (A) displays the variation in the percentage of forests composed of cores and edges over time. Section (B) shows changes in the area (ha) of cores and edges. The bluish and reddish horizontal dotted lines in Section B represent the average annual areas of cores and edges, respectively.
Forest quality
The amount of older forest cores in large fragments, which may represent the PEC’s higher-quality habitats, was only about 62,058 ha in 2017, representing 12% of the total forest cover in that year (Table 1 and S6 Fig).
Discussion
Our main finding is that a large portion of the PEC forests is threatened not only by fragmentation and edge effects but also by forest rejuvenation, with older forest cores in large fragments representing only 12% (62,058 ha) of the remaining forest cover in 2017. Temporal analyses revealed an overall reduction of around 5% in the total forest cover from 1985 to 2020, and the highest deforestation rates were found between 1985 and the early 2000s. At least 87% of the 4.4 Mha of the original PEC forest cover was devastated before 1985, suggesting that most of its fragments may have been isolated for many decades or centuries. Although we observed a tendency for forest cover recuperation in the last two decades, it was insufficient to compensate for the losses that occurred in the decades of 1980 and 1990. Our estimate of the total remaining forest cover of the PEC was higher than previously reported, i.e., in 1990–1995 (256,581 ha) [30], 2005 (360,455 ha) [22], and 2001–2007 (322,372 ha) [29]. We suggest that it has occurred due to: i) the recuperation of part of the forests during the last decade, ii) differences in the data spatial resolution, iii) differences in the minimum size of fragments considered as forests, and iv) the use of more conservative methods for forest cover classification in the abovementioned studies. The situation observed in the PEC may also extend to other tropical forests in human-modified landscapes, and our results suggest the need for further investigations to characterize the degradation of these megadiverse ecosystems around the globe.
We observed a positive balance between forest regeneration and deforestation and a decrease in forest fragmentation in the last decade. This tendency may be related to the PEC’s ongoing initiatives of forest restoration, linked to the Atlantic Forest Restoration Pact [23] and/or changes in practices of the sugar-cane industry in the face of the increasing need for producing “environmentally correct” products since the 1990s [24]. However, this should not obscure the dramatic conservation status of the PEC forests. As expected, the drastic deforestation of the decades of 1980 and 1990, and the tendency of forest regeneration registered during the last decade, resulted in a process of forest rejuvenation about 2.5 times higher than that estimated for the whole Brazilian Atlantic Forest [16]. Despite the importance of secondary forests in maintaining a fraction of the original biological diversity in certain regions [51–53], younger forests may not maintain high-quality habitats, and their ecological communities can be altered [54–56]. In the PEC, areas that have been through restoration programs, for instance, were only a third as dense as older forest remnants and maintained considerably different tree communities, with only half of the original vegetal species richness [57].
It is also of great concern that the PEC forests are currently mainly distributed (roughly 90%) into very small fragments (< 10 ha) and that the average fragment size is only about 11 ha. This scenario is more pessimistic than previous reports that informed that 73.3% of the PEC fragments were smaller than 10 ha [29]. Our findings were also more alarming than the previous estimates for the entire Brazilian Atlantic Forest, in which 80% of fragments were smaller than 50 ha [22]. The mean fragment size of the PEC was also smaller than the overall values observed for tropical forests (17 ha for the Americas and 13 ha for Asia and Australia) [58].
Nowadays, only about 12% (62,058 ha) of the PEC forest cover is composed of higher-quality habitats (i.e., older forest cores in fragments larger than 1,000 ha). It is worth noting, however, that we considered as older, the forests present in 1985 not removed until 2017, meaning that we have not discriminated between the areas targeted to selective logging and those that could have regenerated just before 1985. Then, the amount of primary areas is certainly smaller than our estimate. This may be the reason why old-growth forests with full capacity of carbon storage represent only 8% of the total forest cover of the PEC [59, 60]. Furthermore, we only evaluated edges effects within the first 50 m from the borders. However, forest-dwelling and niche specialized tropical species tend to be highly sensitive to edge effects, which can extend up to 400 m into the interior of tropical forests [61–66].
The scenario of intense forest rejuvenation and fragmentation highlights the importance of maintaining the last older core areas in larger fragments of the PEC to preserve taxa dependent on older forests to thrive and to serve as sources of biodiversity for the regenerating areas. We suggest that ensuring effective protection of these larger blocks of older forest cores and increasing connectivity to their surrounding forests must be top priorities for the conservation of the PEC forest-dependent biodiversity. The increase in metrics related to habitat quality over the last decades may indicate a slight recovery of the native forest cover in the region, but forest rejuvenation, fragmentation, and edge effects are still undergoing threats to the PEC. Our results evidenced that forest cover information alone may provide a false scenario about the conservation status of the Brazilian Atlantic Forest. By addressing the temporal component and investigating the spatial characteristics of the fragments, we provide a more realistic scenario of the PEC’s forest cover and more precise information for conservation practitioners and decision-makers.
Supporting information
S1 Fig. Moving window temporal filter for classifying deforestation.
https://doi.org/10.1371/journal.pone.0291234.s001
(TIF)
S2 Fig. Moving window temporal filter for classifying forest regeneration.
https://doi.org/10.1371/journal.pone.0291234.s002
(TIF)
S3 Fig. Scheme of the methods for the classification of deforestation according to fragment size.
https://doi.org/10.1371/journal.pone.0291234.s003
(TIF)
S4 Fig. Scheme for the classification of older forests deforestation.
https://doi.org/10.1371/journal.pone.0291234.s004
(TIF)
S5 Fig. Older and younger forests of the Pernambuco Endemism Center.
The current distribution of older and younger forests according to (A) the municipalities of the PEC, and (B) the spatial distribution and configuration of older and younger forests. The layer of administrative boundaries used in this map was obtained from the open-access dataset provided by the Brazilian Institute of Geography and Statistics (IBGE).
https://doi.org/10.1371/journal.pone.0291234.s005
(TIF)
S6 Fig. Quality of the Atlantic Forest cover over the Pernambuco Endemism Center.
Current (2017) classification of forests according to age and edge effects for (A) large (> 1,000 ha), and (B) medium (100–1,000 ha) fragments over the Pernambuco Endemism Center.
https://doi.org/10.1371/journal.pone.0291234.s006
(TIF)
S1 Table. Landscape metrics of forests over the Pernambuco Endemism Center.
Class-level metrics related to forest cover area (total and by four classes of fragment size), number of fragments (total and by four classes of fragment size), largest fragment area, mean fragment area, core and edge areas. Values presented in hectares for area-related metrics.
https://doi.org/10.1371/journal.pone.0291234.s008
(DOCX)
S2 Table. Deforestation and forest regeneration over the Pernambuco Endemism Center.
Deforestation was classified according to fragment size (deforestation of very small, small, medium, and large fragments) and forest age (in deforestation of older and younger forests). Values presented in hectares.
https://doi.org/10.1371/journal.pone.0291234.s009
(DOCX)
S3 Table. Breakpoints and linear trends of forest cover area and number of fragments of the Pernambuco Endemism Center.
Forests were classified according to fragment size (very small, small, medium, and large fragments), and metrics were calculated separately for all of them and total. Values for areas presented in hectares.
https://doi.org/10.1371/journal.pone.0291234.s010
(DOCX)
References
- 1. Hoang NT, Kanemoto K. Mapping the deforestation footprint of nations reveals growing threat to tropical forests. Nat Ecol Evol. 2021;5: 845–853. pmid:33782576
- 2. Roberts P, Hamilton R, Piperno DR. Tropical forests as key sites of the “Anthropocene”: Past and present perspectives. Proceedings of the National Academy of Sciences. 2021;118: e2109243118. pmid:34580229
- 3. Keenan RJ, Reams GA, Achard F, de Freitas J V, Grainger A, Lindquist E. Dynamics of global forest area: Results from the FAO Global Forest Resources Assessment 2015. For Ecol Manage. 2015;352: 9–20.
- 4. Tapia-Armijos MF, Homeier J, Espinosa CI, Leuschner C, de la Cruz M. Deforestation and Forest Fragmentation in South Ecuador since the 1970s –Losing a Hotspot of Biodiversity. PLoS One. 2015;10: e0133701. pmid:26332681
- 5. Laurance WF, Camargo JLC, Luizão RCC, Laurance SG, Pimm SL, Bruna EM, et al. The fate of Amazonian forest fragments: A 32-year investigation. Biol Conserv. 2011;144: 56–67.
- 6. Gonzalez A, Germain RM, Srivastava DS, Filotas E, Dee LE, Gravel D, et al. Scaling-up biodiversity-ecosystem functioning research. Ecol Lett. 2020;23: 757–776. pmid:31997566
- 7. Liu J, Wilson M, Hu G, Liu J, Wu J, Yu M. How does habitat fragmentation affect the biodiversity and ecosystem functioning relationship? Landsc Ecol. 2018;33: 341–352.
- 8. Costa-Araújo R, Regolin AL, Martello F, Souza-Alves JP, Hrbek T, Ribeiro MC. Occurrence and conservation of the vulnerable titi monkey Callicebus melanochir in fragmented landscapes of the Atlantic forest hotspot. Oryx. 2021;55: 916–923.
- 9. Galetti M, Giacomini HC, Bueno RS, Bernardo CSS, Marques RM, Bovendorp RS, et al. Priority areas for the conservation of Atlantic forest large mammals. Biol Conserv. 2009;142: 1229–1241.
- 10. Hansen MC, Wang L, Song X-P, Tyukavina A, Turubanova S, Potapov P V, et al. The fate of tropical forest fragments. Sci Adv. 2023;6: eaax8574. pmid:32195340
- 11. Edwards DP, Socolar JB, Mills SC, Burivalova Z, Koh LP, Wilcove DS. Conservation of Tropical Forests in the Anthropocene. Curr Biol. 2019;29: R1008–R1020. pmid:31593660
- 12. Regolin AL, Cherem JJ, Graipel ME, Bogoni JA, Ribeiro JW, Vancine MH, et al. Forest cover influences occurrence of mammalian carnivores within Brazilian Atlantic Forest. J Mammal. 2017;98: 1721–1731.
- 13. Crooks KR, Burdett CL, Theobald DM, King SRB, Di Marco M, Rondinini C, et al. Quantification of habitat fragmentation reveals extinction risk in terrestrial mammals. Proc Natl Acad Sci. 2017;114: 7635–7640. pmid:28673992
- 14. Mohammadi A, Almasieh K, Nayeri D, Ataei F, Khani A, López-Bao J V, et al. Identifying priority core habitats and corridors for effective conservation of brown bears in Iran. Sci Rep. 2021;11: 1–13. pmid:33441776
- 15. Zhang Z, Sherman R, Yang Z, Wu R, Wang W, Yin M, et al. Integrating a participatory process with a GIS-based multi-criteria decision analysis for protected area zoning in China. J Nat Conserv. 2013;21: 225–240.
- 16. Rosa MR, Brancalion PHS, Crouzeilles R, Tambosi LR, Piffer PR, Lenti FEB, et al. Hidden destruction of older forests threatens Brazil’s Atlantic Forest and challenges restoration programs. Sci Adv. 2021;7: eabc4547. pmid:33523918
- 17. Bürgi M, Hersperger AM, Schneeberger N. Driving forces of landscape change—current and new directions. Landsc Ecol. 2005;19: 857–868.
- 18. Martensen AC, Saura S, Fortin M-J. Spatio-temporal connectivity: assessing the amount of reachable habitat in dynamic landscapes. Methods Ecol Evol. 2017;8: 1253–1264.
- 19. Ranta P, Blom TOM, Niemela J, Joensuu E, Siitonen M. The fragmented Atlantic rain forest of Brazil: size, shape and distribution of forest fragments. Biodivers Conserv. 1998;7: 385–403.
- 20. Metzger JP, Sodhi N. Conservation issues in the Brazilian Atlantic forest. Biol Conserv. 2009;142: 1138–1140.
- 21.
Mittermeier RA, Turner WR, Larsen FW, Brooks TM, Gascon C. Global biodiversity conservation: the critical role of hotspots. In: Zachos FE, Habel JC. Biodiversity hotspots: distribution and protection of conservation priority areas. Germany: Springer; 2011; p. 3–22.
- 22. Ribeiro MC, Metzger JP, Martensen AC, Ponzoni FJ, Hirota MM. The Brazilian Atlantic Forest: How much is left, and how is the remaining forest distributed? Implications for conservation. Biol Conserv. 2009;142: 1141–1153.
- 23.
Lins-e-Silva ACB, Ferreira PSM, Rodal MJN. The north-eastern Atlantic Forest: biogeographical, historical, and current aspects in the sugarcane zone. In: Marques MCM, Grelle CEV. The Atlantic Forest: History, Biodiversity, Threats and Opportunities of the Mega-diverse Forest. Switzerland: Springer; 2021; p. 45–61.
- 24. Tabarelli M, Roda SA. An opportunity for the Pernambuco Endemism Center. Nat Conserv. 2005;3: 128–134.
- 25. Barnett JM, Buzzetti DRC. A new species of Cichlocolaptes Reichenbach 1853 (Furnariidae), the ‘gritador-do-nordeste’, an undescribed trace of the fading bird life of northeastern Brazil. Rev Bras Ornitol. 2014;22: 75–94.
- 26. Pereira GA, Dantas S de M, Silveira LF, Roda SA, Albano C, Sonntag FA, et al. Status of the globally threatened forest birds of northeast Brazil. Pap Avulsos Zool. 2014;54: 177–194.
- 27.
Roda SA, Pereira GA, Albano C. Conservação de Aves endêmicas e ameaçadas do Centro de Endemismo Pernambuco. Recife: Editora Universitária UFPE; 2011.
- 28. Silveira LF, Olmos F, Long AJ. Birds in Atlantic Forest fragments in north-east Brazil. News Rev Feat. 2003;1832: 32–46.
- 29. Mendes-Pontes AR, Beltrão ACM, Normande IC, Malta A de JR, Silva Júnior AP da, Santos AMM. Mass extinction and the disappearance of unknown mammal species: scenario and perspectives of a biodiversity hotspot’s hotspot. PLoS One. 2016;11: e0150887. pmid:27191719
- 30.
MMA. Avaliação e ações prioritárias para a conservação da biodiversidade da Mata Atlântica e Campos Sulinos. Brazil: Ministério do Meio Ambiente; 2000.
- 31. Gorelick N, Hancher M, Dixon M, Ilyushchenko S, Thau D, Moore R. Google Earth Engine: Planetary-scale geospatial analysis for everyone. Remote Sens Environ. 2017;202: 18–27.
- 32. MapBiomas. Collection 6 of the Annual Series of Land Use and Land Cover Maps of Brazil. 2022 [cited 27 Mar 2023]. Available from: http://mapbiomas.org
- 33. Nemésio A, Santos Junior JE. Is the “Centro de Endemismo Pernambuco” a biodiversity hotspot for orchid bees? Braz J Biol. 2014;74: S078–S092. pmid:25627369
- 34. Alvares CA, Stape JL, Sentelhas PC, Gonçalves JL de M, Sparovek G. Köppen’s climate classification map for Brazil. Meteorol Zeitschrift. 2013;22: 711–728.
- 35.
Wege DC, Long AJ. Key areas for threatened bird in the neotropics. Cambridge: BirdLife International; 1995.
- 36.
Stattersfield AJ. Endemic bird areas of the world: priorities for biodiversity conservation. Cambridge: BirdLife International; 1998.
- 37. Garbino GST, Rezende GC, Fernandes–Ferreira H, Feijó A. Reconsidering mammal extinctions in the Pernambuco Endemism Center of the Brazilian Atlantic Forest. Anim Biodiv Conserv. 2018;41: 69–77.
- 38. Mendes-Pontes AR, Mariz-Beltrão AC, Melo-Santos AM. Reconsidering mammal extinctions in the Pernambuco endemism center of the Brazilian Atlantic Forest: a critique. Anim Biodivers Conserv. 2019;42.1: 69–77.
- 39.
ICMBio. Livro Vermelho da Fauna Brasileira Ameaçada De Extinção: volume III–Aves. Brazil: Ministério do Meio Ambiente; 2018.
- 40. Butchart SHM, Lowe S, Martin RW, Symes A, Westrip JRS, Wheatley H. Which bird species have gone extinct? A novel quantitative classification approach. Biol Conserv. 2018;227: 9–18.
- 41.
FAO. Global Forest Resources Assessment 2020: Terms and Definitions. Italy: Food and Agriculture Organization of the United Nations; 2020.
- 42.
Lloyd C. Spatial data analysis: an introduction for GIS users. United Kingdom: Oxford University Press; 2010.
- 43. Dutrieux L, DeVries B, Verbesselt J. Utilities to monitor for change on satellite image time-series. 2014 [cited 27 Mar 2023]. Available from: https://github/com/dutri001/bfastSpatial
- 44. R Core Team. A language and environment for statistical computing. 2009 [cited 27 Mar 2023]. Available from: https://www/R-project/org
- 45. Muggeo VMR. Segmented: an R package to fit regression models with broken-line relationships. R news. 2008;8: 20–25.
- 46. Taddeo S, Dronova I. Landscape metrics of post-restoration vegetation dynamics in wetland ecosystems. Landsc Ecol. 2020;35: 275–292.
- 47. Nanni AS, Sloan S, Aide TM, Graesser J, Edwards D, Grau HR. The neotropical reforestation hotspots: A biophysical and socioeconomic typology of contemporary forest expansion. Glob Environ Change. 2019;54: 148–159.
- 48. Melo FPL, Dirzo R, Tabarelli M. Biased seed rain in forest edges: evidence from the Brazilian Atlantic forest. Biol Conserv. 2006;132: 50–60.
- 49. Melo FPL, Lemire D, Tabarelli M. Extirpation of large-seeded seedlings from the edge of a large Brazilian Atlantic forest fragment. Ecoscience. 2007;14: 124–129.
- 50. Silva MAM, Alencar PGAM, Guerra TNF, Melo AL, Lins ACB, Rodal MJN. Edge effects on the structure and dynamics of an Atlantic Forest fragment in northeastern Brazil. Brazil J Agric Sci. 2015;10: 538–543.
- 51. Chazdon RL, Peres CA, Dent D, Sheil D, Lugo AE, Lamb D, et al. The potential for species conservation in tropical secondary forests. Conserv Biol. 2009;23: 1406–1417. pmid:20078641
- 52. Dent DH, Wright SJ. The future of tropical species in secondary forests: a quantitative review. Biol Conserv. 2009;142: 2833–2843.
- 53. Metzger JP, Martensen AC, Dixo M, Bernacci LC, Ribeiro MC, Teixeira AMG, et al. Time-lag in biological responses to landscape changes in a highly dynamic Atlantic forest region. Biol Conserv. 2009;142: 1166–1177.
- 54. Bihn JH, Verhaagh M, Brändle M, Brandl R. Do secondary forests act as refuges for old growth forest animals? Recovery of ant diversity in the Atlantic forest of Brazil. Biol Conserv. 2008;141: 733–743.
- 55. Santos BA, Peres CA, Oliveira MA, Grillo A, Alves-Costa CP, Tabarelli M. Drastic erosion in functional attributes of tree assemblages in Atlantic forest fragments of northeastern Brazil. Biol Conserv. 2008;141: 249–260.
- 56. Pinho BX, Melo FPL, Arroyo‐Rodríguez V, Pierce S, Lohbeck M, Tabarelli M. Soil‐mediated filtering organizes tree assemblages in regenerating tropical forests. J Ecol. 2018;106: 137–147.
- 57. Santos-Costa A, Malhado ACM, Bragagnolo C, Correia RA, Ladle RJ. Ecological outcomes of Atlantic Forest restoration initiatives by sugar cane producers. Land Use Policy. 2016;52: 345–352.
- 58. Taubert F, Fischer R, Groeneveld J, Lehmann S, Müller MS, Rödig E, et al. Global patterns of tropical forest fragmentation. Nature. 2018;554: 519–522. pmid:29443966
- 59. Paula MD, Costa CPA, Tabarelli M. Carbon storage in a fragmented landscape of Atlantic forest: the role played by edge-affected habitats and emergent trees. Trop Conserv Sci. 2011;4: 349–358.
- 60. Dias TC, Silveira LF, Pironkova ZI, Francisco MR. Greening and browning trends in a tropical forest hotspot: Accounting for fragment size and vegetation indices. Remote Sens Appl. 2022;26: 100751.
- 61. Murcia C. Edge effects in fragmented forests: implications for conservation. Trends Ecol Evol. 1995;10: 58–62. pmid:21236953
- 62. Laurance WF, Nascimento HEM, Laurance SG, Andrade A, Ewers RM, Harms KE, et al. Habitat fragmentation, variable edge effects, and the landscape-divergence hypothesis. PLoS One. 2007;2: e1017. pmid:17925865
- 63. Laurance WF. Theory meets reality: how habitat fragmentation research has transcended island biogeographic theory. Biol Conserv. 2008;141: 1731–1744.
- 64. Hansbauer MM, Storch I, Leu S, Nieto-Holguin J-P, Pimentel RG, Knauer F, et al. Movements of neotropical understory passerines affected by anthropogenic forest edges in the Brazilian Atlantic rainforest. Biol Conserv. 2008;141: 782–791.
- 65. Lopes AV, Girão LC, Santos BA, Peres CA, Tabarelli M. Long-term erosion of tree reproductive trait diversity in edge-dominated Atlantic forest fragments. Biol Conserv. 2009;142: 1154–1165.
- 66. Pfeifer M, Lefebvre V, Peres CA, Banks-Leite C, Wearn OR, Marsh CJ, et al. Creation of forest edges has a global impact on forest vertebrates. Nature. 2017;551: 187–191. pmid:29088701