UC Berkeley

Tropical forests throughout the world are rapidly being converted to agriculture. Remaining forests are often fragmented, threatening area-demanding species, such as apex predators and mega-herbivores (e.g. elephants). The loss of predators can trigger trophic cascades, whereby prey species increase in abundance, altering food web dynamics. Fragmentation also increases hunters’ and poachers’ access to previously remote areas, adding an additional threat to megafauna. A pressing challenge in conservation biology is to understand where, why, and over what time scales these unintended secondary processes are degrading remaining forests. My dissertation seeks to address this challenge by exploring how forest loss, predator loss, hunting, and rapid oil palm agricultural expansion are affecting tropical forest floral and faunal communities in Southeast Asia.

My first chapter, an introduction, discusses the process and theories on how land use change affects species and drives ecological cascades. In my second chapter, I grapple with the approaches used to study the impacts of agricultural expansion on biodiversity, and how different methods can shape outcomes for conservation planning. In this chapter, I explore the land sparing versus land sharing framework for conservation planning. Land sparing advocates meeting production targets through increasing yields on existing farmland (intensification) in order to protect natural areas from further conversion (set asides). Land sharing promotes conservation within agricultural landscapes using wildlife-friendly farming practices. Using a literature review, I argue that there is an emerging consensus of ecological theory, empirical data, and direct case studies supporting the “land sparing” approach to conserve biodiversity in tropical forested landscapes. Studies of land sharing landscapes routinely report the complete absence of over 50% of forest species, even in wildlife-friendly agroforestry systems. This indicates most species are sensitive to any farming activities, and so conservation should focus on minimizing forest loss. I conclude by exploring three important considerations to effectively implement a land sparing strategy: (i) the tradeoffs and synergies with other ecosystem services (e.g. carbon, water quality); (ii) the economic benefits of forest offsets (e.g. from REDD+ carbon payments) and; (iii) the sociopolitical obstacles to insuring effective forest offsets are created.

In the third chapter, I turn to the abiotic impacts of agricultural expansion, comparing and contrasting the microclimate of agricultural lands and forest. Specifically, I examine how abiotic conditions in oil palm plantations vary throughout the life cycle of a plantation. I present a chronosequence study on microclimate and vegetation structure in protected forest and its surrounding oil palm plantations surrounding in Peninsular Malaysia. My results indicate that understory vegetation is twice as tall in young plantations, but leaf litter depth and total epiphyte abundance is twice as high in old plantations. Plantations are also substantially hotter (+2.84 C) and drier (+0.80hPa vapor pressure deficit) than forests during the day, but there are no nocturnal microclimate differences between the two. These findings are important to revealing the habitat heterogeneity throughout the 25-year plantation lifecycle. Based on these results, I develop environmental guidelines to improve the spatio-temporal planning of oil palm plantations for wildlife, drawing heavily on timber management practices. I conclude that oil palm plantations can be developed to create a permeable matrix in order to connect remaining forests habitats and increase the conservation value of the landscape.

In Chapter 4, I explore how agricultural expansion is affecting faunal communities, focusing on predators. Accurately monitoring predator populations is an essential but difficult challenge for conservation biologists. Until very recently, many of the methods employed by conservationists overestimated population densities and sizes. This is particularly true for apex predators like tigers, which are important to maintaining ecosystem function and are especially sensitive to habitat loss and poaching. To accurately and defensibly estimate current Sumatran tiger populations, in this chapter, I present new data collected from camera trap arrays across three expansive landscapes (843-999km2), which together make up the UNESCO Tropical Rainforest World Heritage Sites of Sumatra. I use these data to estimate tiger densities using the spatial-capture-recapture approach. Then, to compare my results with other studies, I develop a new approach to correct for biases in previous research and conduct a meta-analyses to draw inferences about the determinants of densities. I find that traditional mark-recapture techniques published before 2010 inflated estimates of tiger densities by 63.3% on average. Controlling for this bias, tiger densities increased 4.2%/yr from 1999 to 2014 and primary forest densities were 58% higher than disturbed forests. Based on my study, I estimate there are 734 ± 304 tigers within or connected to source landscapes. These results highlight that Sumatran tiger densities in remaining forest may be increasing. However, forest loss, fragmentation, and poaching have reduced the total population and threaten the subspecies with extinction.

In Chapter 5, I delve deeper into the impacts of agricultural expansion by focusing on changes in faunal and floral community dynamics. There has been significant debate over whether herbivores are regulated from the “top-down” by predators or from the “bottom up” as plants adapt and limit herbivores’ consumption. In altered ecosystems, such as forests fragmented by agriculture expansion, the widespread loss of predators is thought to be the main destabilizing force. I challenge this status quo by examining the effect of bottom-up agricultural resource subsidies (palm oil fruits) in controlling forest faunal populations as well as the effect on tree communities. To determine these impacts, my fifth chapter takes advantage of a landscape-scale manipulation of agricultural resources adjacent to a 2500 ha forest reserve in Peninsular Malaysia. First, I evaluate whether resource subsidies affect forest wildlife populations by compiling two decades of wild boar (Sus scrofa, a crop-raiding species) abundance data at the Pasoh Research Forest in Peninsular Malaysia. Second, I evaluate if altered wild boar abundance had cascading impacts on the vegetation community by re-censusing seven wildlife exclosures established in 1996. Finally, I evaluate how wild boar predation shaped tree sapling diversity using 24 years of tree census data from the Smithsonian Institute’s Center for Tropical Forest Science 50-ha research plot at the Pasoh. I find that the forest density of wild boar was 100 times higher when palm oil plantations were fruiting (1996-2000 and 2007-2014) than when plantations were cleared and fruit was unavailable (2001-2006). Second, wild boars’ predation of tree saplings (100-200 cm in height) led to a to 46% reduction in the density of 1-2cm dbh saplings over a 23-year period. Third, while all species of tree saplings had lower abundances, there was a 13% increase in tree diversity as measured by the Fisher’s alpha diversity index. This suggests wild boars exhibit density dependent selective mortality on trees that disproportionately reduces the abundance of common tree species.

My final chapter turns to the role of humans in these new forest-plantation landscapes. In this chapter, I examine how the immigration of farmers alters hunting practices. Using information from in-depth interviews with hunters, agricultural workers, and wild meat dealers in Jambi province, Sumatra, I first describe how plantations have affected local human demography. Then, I explore how wildlife hunting and consumption rates vary between different indigenous and immigrant ethnic groups. I also uncover how hunting near palm oil plantations has become primarily a commercial endeavor for managing crop-raiding wild boars. These results also indicate that wild boars may be experience ecological release, either from the loss of predation by tigers, or increased food available in agricultural fields. I discuss how proper management could reduce crop damage and also yield large amounts of wild boar meat with relatively little by-catch of threatened wildlife.

Taken together, these studies indicate that oil palm habitat is unsuitable for most forest species, and that forest fragments in oil palm landscapes are undergoing a loss of predators and thus potentially suffer from trophic cascades and subsidy cascades. I found that cross-border agricultural resource subsidies shape wildlife communities and devastate tree sapling communities. These results also suggest that predator loss and agricultural expansion can lead to combined trophic and subsidy cascades with heightened effects, and that this form of indirect forest degradation may be widespread in the region and globally. As such, my results indicate that protecting large continuous forests is necessary to preserve functioning Southeast Asian food webs. This conclusion has direct applicability across the tropical rainforests regions where oil palm agriculture is rapidly expanding, and more generally informs the discussion on how to achieve conservation goals.