Our study provides valuable information on forest dynamics, as well as their impacts and drivers in a landscape within one of the most threatened regions of the Atlantic Forest hotspot. Over the 35 years assessed, there was a replacement of old-growth forests by younger forests. The spatial distribution of forest cover was directly affected by the expansion and contraction of different land uses. The results allowed us to outline a general pattern of forest dynamics, characterised by forest loss, regeneration-driven gains, and forest stability.
The conversion of forests into agriculture and pasture fields, and vice versa, marked the dynamics in this landscape. A large proportion of forest loss occurred due to conversion for agriculture, with a smaller amount attributed to the expansion of urban infrastructure, as documented in previous studies (Teixeira et al. 2009; Rezende et al. 2015; Molin et al. 2017; Lins-e-Silva et al. 2021; Rosa et al. 2021). Prior research also indicated that the loss of Atlantic Forest was driven by the slope in this hotspot, particularly in flat areas and gentle slopes (Teixeira et al. 2009; Rezende et al. 2015; Molin et al. 2017), due to the ease of transportation and crop mechanisation (Rezende et al. 2015). Forest regeneration in the studied landscape primarily occurred in areas that were previously used for agriculture but that have now been abandoned. In fact, in the Atlantic Forest, regeneration is explained by the abandoned area, particularly of those areas with high slopes, due to the impracticability of mechanised agriculture (Molin et al. 2017). Slope, in general, is a decisive factor for vegetation regeneration in the Atlantic Forest (Teixeira et al. 2009; Rezende et al. 2015; Molin et al. 2017). However, in the present study and in the study conducted by Molin et al. (2018), slope was not a significant driver, as it appears to be irrelevant for regeneration gains in a landscape with forest cover above 50%.
Our results indicated that proximity to water networks is another key explanatory variable for forest loss and gain. However, while more than half of the lost fragments were located in riparian areas in the studied landscape, this may not always be the pattern in other forest landscapes, and forest loss can also be observed as one moves away from rivers (Teixeira et al. 2009). This discrepancy can be explained by previous research which stated that sugarcane was the prominent crop in riparian areas of the Atlantic Forest in Northeast and Southeast Brazil until the 2000s (Gunkel et al. 2007; Silva et al. 2007). However, in the studied region, sugarcane cultivation did not expand in the past two decades, and forest conversion likely occurred as a result of urban infrastructure expansion. As revealed in previous research, forest regeneration was also driven by proximity to water networks (Teixeira et al. 2009; Molin et al. 2017; Rosa et al. 2021). Additionally, it was observed that in the current landscape, 90% of the medium-sized fragments and 56% of the small fragments are in riparian areas. Therefore, the gain in forest regeneration may have occurred due to the abandonment of riparian sites, as a result of compliance with laws protecting native vegetation, such as the Atlantic Forest Law (Brasil 2006), which prohibits the conversion of forests for other land uses, and the Law for the Protection of Native Vegetation (Brasil 2012), the Environmental Crimes Law (Brasil 1998), and the State Law that delimits Protection Areas for Water Sources (Pernambuco 1986), combined with the high resilience capacity of these areas (Chazdon and Guariguata, 2016).
Riparian areas provide privileged water conditions for aboveground biomass and carbon storage, contributing to climate change mitigation (Brancalion et al. 2016; Brancalion et al. 2021). Those areas also contribute to connectivity between forest patches and ensure water security (Ferreira et al. 2015). Consequently, the loss of riparian forests leads to significant changes in ecosystem functioning. Hence, if the future goal is the conservation and increase of ecosystem services, legislation and enforcement must be further encouraged, as several countries, including Brazil, have recently committed to halting and reversing forest loss and degradation by 2030, as they signed the Glasgow Climate Pact (Gasser et al. 2022).
The forest gain in protected areas validated the effectiveness of protection measures in these areas, unlike the observed decrease in forest cover in protected areas of the Atlantic Forest in Paraguay (Ponte et al. 2017). Currently, only 10% of the remaining vegetation cover of the Brazilian Atlantic Forest is within protected areas (Rezende et al. 2018), which are less effective in protecting this biome than in protecting the Amazon (Sobral-Souza et al. 2018). Therefore, we recommended increasing protected areas in this hotspot with proper governance. Additionally, actions that support forest gain in existing protected areas, such as assisted natural regeneration (Chazdon 2017), should be intensified, as most of the native vegetation cover is located outside protected areas and may face interference due to public interest and social utility (Rezende et al. 2018).
Although it has been revealed that forest cover in the Brazilian Atlantic Forest has remained relatively constant over the past 30 years, this stability masks the destruction of old-growth forests by the increasing initial forest cover gain, as also found by Rosa et al. (2021). This research assessed forest stability as the amount of forest cover that showed no change during the two time intervals. Thus, it was estimated that over 35 years, 44% of the forest remained stable, corresponding to 75% of the forest cover in 2020. However, the study also revealed the existence of a few fragments of old-growth vegetation larger than 30 ha, confirming the spatially hyper-fragmented structure of the Brazilian Atlantic Forest (Ribeiro et al. 2009; Rosa et al. 2021). Furthermore, it was observed that the spatial-temporal stability of forest cover is associated with the locations of adjacent administrative units, which may be directly related to the increased accessibility of environmental agencies responsible for forest monitoring and protection. The proximity of urban areas reflects society's engagement in enforcing environmental laws. Studies assert that the public’s participation in reporting cases of forest suppression or their suspicions of actions that could lead to disturbance and damage, is high and helps to uncover possible crimes in Atlantic Forest areas (Silva et al. 2017; Molin et al. 2017).
Regarding the methodology applied, the minimum threshold of 1 ha applied in previous analyses to aggregate forest gain and loss (Rosa et al. 2021), may have limited the ability to identify the gain and loss of small fragments. For this study, a threshold of 0.5 ha was applied to reduce this uncertainty. Since the analysed time series began in 1985, it was not possible to identify native vegetation patches that regenerated before this date. Therefore, some forests which were counted as a loss in the first interval may have been composed of young forests. Despite this potential methodological limitation, the data allowed for the description and explanation of landscape changes and their drivers. To reduce this uncertainty, we analysed and classified the 'loss of old-growth forest' to include only the forest loss that occurred in the second period, i.e., after 2003. Thus, regardless of the net forest gain over 35 years, biodiversity has been negatively impacted because younger forests do not compensate for the ongoing loss of old-growth forests (Rosa et al. 2021), resulting in a forest mosaic with different successional stages.
We know that high percentages of forest cover are necessary for the conservation of biodiversity and ecosystem services. However, if only the net gain of forest is considered, in light of the Brazilians commitments to forest restoration such as the Nationally Determined Contributions under the Paris Agreement (UNFCCC, 2015; Silva Junior et al. 2020; Brasil, 2022), the Pact for the Restoration of the Atlantic Forest (Pinto et al. 2014), the National Plan for Recovery of Native Vegetation (Pinto and Voivodic, 2021), and recently the Glasgow Declaration (Gasser et al. 2022), there will be an increase of millions of hectares in forest cover with few functional benefits and even a reduction in species conservation and ecosystem services (Rosa et al. 2021). Mature forests are essential for conserving tropical biodiversity due to the inability of many species to recolonise younger forests (Gibson et al. 2011; Rosa et al. 2021). Species rely on older and less disturbed habitats to persist in a highly anthropized and fragmented landscape (Gibson et al. 2011).