Disentangling island food-web connections in Guam after the brown treesnake (Boiga irregularis) introduction

Abstract

Shifts in predator abundance or identity strongly affects prey populations. If predators decline and predation pressures are relaxed, prey populations can increase, and the plants they feed on may experience more herbivory in a trophic cascade. Additionally, behaviors prey uses to escape from predators may decline. After an apex predator (brown treesnake [Boiga irregularis]) was introduced to Guam in the Pacific in the 1940's, forest birds that had previously been the top predators of arthropods and other small vertebrates were functionally extirpated. In this dissertation, I evaluate whether the loss of birds has led to a tropic cascade in which arthropods increase in abundance, leading to increased herbivory on plants. I quantify the effects of the loss of birds on arthropods and plants in Chapter 1, I explore shifts in predation in Chapter 2 and measure shifts in anti-predator behaviors in Chapter 3. My overarching goal is to deepen our understanding of how removal of former top predators influences trophic ecology, behavior and community structure. In Chapter 1, I compared total number of arthropods across all species and herbivory between Guam and the nearby islands of Saipan and Rota, which retain their native bird populations. I hypothesized that loss of birds would lead to an increased abundance of arthropods in Guam but instead found that Guam had the lowest arthropod abundances of the three islands. Contrary to a previous study, examining separate groups of arthropods revealed a consistent pattern for arachnids, with Guam having the fewest. In contrast, there were more ant species and ant presence found in Rota. In accordance with the data on arthropod abundance, I found that plants experienced less herbivory on Guam. The overall reduced arthropod abundance in Guam may explain the reduced number of arachnids and herbivory that the island experienced. The presence of top predators, such as insectivorous birds, on the other islands may have maintained arthropod predators. My study suggests that the loss of birds through the introduction of the brown treesnake may have altered the abundance and herbivory on plants. I did not observe a classic trophic cascade as expected. In Chapter 2, I investigated the indirect trophic impacts of invasive snakes on arthropods. I hypothesized Guam, without birds, would have reduced predation pressure on arthropods compared to the nearby islands of Rota and Saipan, which have birds. To evaluate my hypothesis, I measured attack rates on model caterpillars, combining data I collected in 2022 with data from collaborators collected in 2013. In both years, attack rates from all possible predators (e.g. birds, arthropods, etc.) on caterpillars were greatest in islands with birds present, and lowest in Guam where no birds were present. However, from identifying the identity of the predators, differences in attack rates were driven by arthropods and lizards rather than birds. On Guam, I observed high attack rates by arthropods. The suppression of birds and small reptiles by snakes have been linked to the release of predacious arthropods. Contrary to smaller-scale experimental studies that reveal the release of herbivorous arthropods when birds are excluded, I found that birds played a minimal role in predating caterpillars on all islands. Thus, my results suggest that reptiles may be underappreciated predators in tropical forest food webs. Finally, in Chapter 3, I ask whether the loss of insectivorous birds from Guam has changed the abundance and anti-predator behaviors of butterflies. I addressed this question by comparing butterfly abundance and behaviors in Guam to those in nearby, similar islands (Rota and Saipan) that retain birds. Butterfly abundance varied by taxon, but they were generally less abundant in islands with birds than in Guam. Behaviors of a common butterfly (Euploia eunice) that I hypothesized to facilitate escape from predators were reduced in Guam compared to islands with birds. Specifically, when butterflies were approached with a model predator (a bird), I could get closer to butterflies before they took flight on Guam than on Saipan. Similarly, on Guam, butterflies exhibited fewer evasive flight behaviors than in Saipan. My study indicates that ecosystem-scale loss of birds dramatically alters both the abundance and behavior of butterflies. These results are important to consider when reintroducing birds to areas where butterflies have lost their defenses against them. In summary, my dissertation found that the loss of birds did not trigger a trophic cascade as expected from literature and bird exclusion studies (Chapter 1). Instead, I found that arthropods and lizards may be key predators in the tropics (Chapter 2). Finally, I found that shifts in butterfly anti-predator behavior have occurred due to the loss of birds (Chapter 3). Through this work, I hope to shine light on the ecological effects of the brown treesnake on the island of Guam and to inspire the people of Guam to write their own scientific narratives.