The fate of forests and its consequences for ecosystem services provision in the Brazilian Atlantic Forest

Although deforestation remains widespread in the tropics, many places are now experiencing significant forest recovery, offering an optimistic outlook for natural ecosystem recovery and carbon sequestration. Natural forest regeneration is a key component of global ecosystem restoration scenarios. Regenerated forests, however, may not persist so a more nuanced understanding of the drivers of forest persistence in the tropics is critical to ensure the success reforestation efforts and carbon sequestration targets. Furthermore, the maintenance of native forests, including young second-growth ones, is essential for the continuous provision of a myriad of ecosystem services that we, as a society, rely on. More specifically, native forests play a crucial role in watershed protection and forest cover loss via changes in land use can lead to deterioration of water quality. Ensuring a sufficient and adequate supply of water for humans and ecosystems is a pressing environmental challenge and land use decisions can severely degrade stream water quality and compromise water supply. This dissertation focusses on two pressing current issues, the dynamics of tropical forest regeneration and the effects of land use on water resources. First, I use a long-term series of detailed land cover data to study forest cover trajectories and persistence of regenerated forest in the Brazilian Atlantic Forest (AF), a restoration hotspot. Secondly, I use 20 years of stream water quality data combined with land cover information to investigate the effects of land cover composition on water resources in the state of São Paulo, Brazil. In Chapter 1, I investigate forest cover trajectories in 3,014 municipalities and quantify the carbon sequestration potential of forest regeneration in the AF. I found that deforestation reversals were the prevalent trend in the region (38%) but concomittant reforestation reversals (13%) suggest that these short-term increases in native forest cover do not necessarily translate into persistent trends, which limited carbon sequestration from reforestation to less than one third of its potential. In Chapter 2, I quantify forest regeneration in the AF and study its persistence. I mapped over 4.47 Mha of native forest regeneration in the region between 1985 and 2019, of which, two thirds persisted until 2019 (3.1 Mha). The relatively low persistence of second-growth forests suggests a rapid turnover of regrowing forests under certain conditions. In Chapter 3, I combine stream water quality data with detailed and land cover information to investigate the effects of landscape composition on the quality of water resources in the state of São Paulo, Brazil. I found that human dominated watersheds had lower overall water quality when compared to conserved ones, with urban cover showing the most detrimental impacts on water quality, while forest cover was associated with a better overall water quality across the studied watersheds. Finally, in Chapter 4, I examine temporal changes in water quality and their association with land use and sewage treatment also in the state of São Paulo. I show that a large proportion of stream water samples failed to meet legal thresholds for at least one water quality metric and that urbanization and agricultural activity led to deterioration of water quality over time, while sewage treatment infrastructure was an important factor in improving water quality. Overall, my dissertation underscores the importance of developing policies that promote second-growth forest persistence to ensure the success of future restoration efforts. It also highlights the need to need to plan and manage landscapes to improve water quality and reduce the growing costs of water treatment, including restoring native forest cover, which is a cost-effective intervention to sustain adequate water quality