The biodiversity and climate impacts of alternative approaches to meeting demands for tropical timber and forest restoration

Alongside agricultural expansion, two other major forces will determine biodiversity and ecosystem services within tropical landscapes of the future. First is timber production, which already exerts substantial impacts on forest species and the climate through selective logging and the development of intensive tree plantations. Second is the restoration of large tracts of degraded lands in order to recover ecosystem services and biodiversity and support livelihoods. Both timber production and restoration will grow dramatically in the near future, as global wood demand grows up to 54% by 2050, and as major restoration agendas gather pace under the UN Decade on Restoration and Bonn Challenge to restore 350 Mha of degraded lands by 2030. Understanding how timber production and restoration can be implemented to navigate trade-offs between competing objectives is critical. In this thesis I consider these trade-offs at different spatial scales, from local to pan-tropical. I do so firstly by advancing the land sparing-sharing framework developed in agriculture to consider the production of wood. I create and evaluate production-matched scenarios that generate timber within a fully forested logging concession, and then across a wider landscape which encompasses the potential for sourcing timber from purpose-grown plantations. Next, I consider how timber production and restoration objectives might intersect over large spatial scales. I predict spatial variation in where tree plantations are likely to expand across Brazil between 2020 and 2030, and evaluate how this expansion may impact areas where ecological restoration would most benefit terrestrial vertebrates. I conclude by considering how restoration can be scaled up across tropical landscapes in ways that are socially and ecologically responsible, and which avoid severe unintended consequences. In Chapter 2, I use previously collected field data to characterise how the abundance of 57 dung beetle species changes with yield along a harvest intensity gradient in a 50,000 ha Brazilian Amazon logging concession. I then use these abundance-yield relationships, combined with company-reported logging rates, to evaluate the biodiversity performance of >8000 hypothetical scenarios, each delivering a fixed volume of wood across the concession. These scenarios follow the sparing-sharing paradigm as originally conceived for farming, but also consider a wide array of mixed approaches. I find that logging approaches that maximise the sparing of old-growth forest (‘extreme sparing’), in exchange for intensifying logging, deliver the highest concession-wide species abundances and population sizes for species negatively affected by harvests. Sparing-style logging also has improved functional diversity outcomes. Just 3% of mixed approaches outperform ‘extreme sparing’ when considering all species, but still involve considerable sparing of old-growth stands combined with higher yields elsewhere than under business-as-usual logging. Yet I also reveal that such concentrated logging is up to 90% less profitable than business-as-usual practices since it limits harvesting of highly-valuable but scattered trees. In Chapter 3, I develop this analytical approach to consider a wider variety of outcomes and harvesting regimes across a 1 Mha landscape in Sabah, Malaysia. I present results from extensive new fieldwork surveys of dung beetles and birds, plus carbon, ‘megatrees’, timber yields, and profits across seven contrasting land uses, including plantations, and spanning >25,000 landscape-level scenarios. I also extend previous analysis by: (i) more comprehensively incorporating post-harvesting trajectories of species’ abundance, and (ii) evaluating how near-optimal harvesting approaches might depend on the composition and degradation status of the starting landscapes. For a wide variety of production targets, biodiversity and climate outcomes would be greatest under production approaches that focus production in plantations alongside reducing logging within either primary or already-logged rainforest ecosystems. This finding holds even as starting landscapes become more degraded, though benefits are greatest within less-degraded landscapes and at higher production targets. However, I again find financial hurdles to sparing-style production: plantation-focused scenarios have much lower harvest profits and also incur greater protection costs in order to conserve forests nolonger falling within active logging concessions. In Chapter 4, I investigate the potential risk of expanding tree plantations into high-value lands forecological restoration in a national-level analysis for Brazil - a timber production powerhouse and restoration hotspot. Using remote-sensing data showing historic patterns of tree plantation expansion, I develop random forest models which use biophysical, economic and spatial-contagion predictor variables to forecast national plantation expansion probability between 2020 and 2030. I then consider the magnitude and extent of potential overlap between areas where future tree plantation expansion is most likely and areas that would deliver the highest restoration benefits for terrestrial vertebrates, if returned to natural habitat cover. I forecast that of 2.8 Mha of future plantation expansion (equivalent to the expansion observed 2010-2020), roughly 1.3 Mha (46%) will occur in the top 30% of restoration priority areas for vertebrates, with the first 1 Mha of expansion having disproportionate impacts. In Chapter 5, I consider how forest restoration can be scaled up and applied effectively in tropicallandscapes across large spatial scales. I review the potential of alternative approaches to restoration in logged forests, abandoned agricultural lands, plantations and farms. I also consider synergies and trade-offs in delivering different restoration objectives. Key environmental challenges to restoration include the risk of displacing agricultural production, the impacts of inappropriate planting in rangelands and savanna woodlands, and the need to embed resilience to climate change and unlock sustainable financing. Overall, I suggest that successful restoration will require addressing issues of poor governance and sociocultural disparities in benefits and costs at landscape levels. In sum, I find that spatially concentrating the production of timber by restricting logging or sourcing more wood from purpose-grown plantations has the potential to substantially improve the biodiversity and climate impacts of meeting future timber demand. However, in order to mitigate the extinction and climate crises this way we need to develop new approaches to supporting such production because it is typically less profitable for harvesters and can also increase overall forest protection costs. When considering the potential benefits of plantations, I also identify the clear need for joined-up approaches to tree plantation expansion which consider alternative land-use options, including the potential to recover species’ populations via targeted ecosystem restoration. More broadly, I find that the potential value of restoration in tropical landscapes is substantial, in part due to the enduring damage caused by historic timber harvests. However, we need integrated, system-level approaches to production and restoration that maximise benefits of different strategies, which avoid interventions that may appear promising locally but cause wider-scale harm to biodiversity and the climate, and which ensure that broadly beneficial landscape-wide management regimes are financially attractive to forestry and plantation enterprises.