Species distribution models and island biogeography: challenges and prospects

Benavides Rios, Eva Consuelo ORCID: 0000-0002-9476-0961 (2024). Species distribution models and island biogeography: challenges and prospects. University of Birmingham. Ph.D.

Preview

BenavidesRios2024PhD.pdf
Text - Accepted Version
Available under License All rights reserved.
Download (15MB) | Preview

Abstract

The distinct characteristics of oceanic islands—restricted geographic boundaries, remoteness, lack of competitors and reduced habitat diversity—have made them global hotspots of endemism. However, these factors also mean that many island species have small populations, low genetic diversity and reduced competitive abilities. The arrival of humans to these systems exposed their fragile vulnerability, causing numerous species extinctions, many more so than on mainland environments.
In addition to their inherent vulnerability, oceanic islands are especially susceptible to environmental change due to their geographic isolation and limited physical space. Unlike many continental species, island species, especially endemics, face restricted opportunities to adapt through migration or by tracking suitable climatic conditions, given their confined distribution and isolation by surrounding aquatic barriers.
To better understand and predict species loss and range shifts in response to global change, this thesis focuses on Species Distribution Models (SDMs). It addresses the limitations of existing SDM frameworks, which are generally designed for continental species with extensive occurrence data, by tailoring the models to the unique conditions of island environments. The thesis focuses on improving SDM applications through the use of high-resolution environmental descriptors (≤500m), given that traditional SDM approaches often rely on coarse climatic variables, which have been shown to be inadequate predictors of island species distributions in many cases. However, the use of high-resolution data raises important questions about how such predictors interact with the often-limited occurrence data available for island species.
To address these challenges, the thesis focused on the Revillagigedo Volcanic Archipelago, using it as a 'model archipelago' to provide insights relevant to a wider range of islands. Plant species were chosen as the focal taxon due to their critical ecological roles and usefulness as conservation indicators, with plant presence/absence data collected through field sampling. The thesis had two main objectives: to enhance SDM applications on islands through the use of high-resolution environmental descriptors, and to assess the robustness of fine-scale predictors, particularly in tropical island environments. The emphasis on tropical island environments was due to both the location of the model archipelago and, more broadly, the rich, yet understudied biodiversity of these regions. This research highlighted the potential of SDMs to deepen our understanding of ecosystems that, despite their global biodiversity significance, is often limited due to insufficient data.
Key findings include the identification of (1) minimum data requirements at very high resolutions (30m) for accurate presence-only species distribution modelling on islands, and (2) strategies to balance data limitations with pseudo-absence data to optimise predictability and accuracy. The research also investigates the impact of spatial inaccuracies in occurrence records, finding that while these inaccuracies are relatively less important for the SDMs of widespread species, they introduce greater uncertainty for range-restricted species; however, these uncertainties were found to be mitigated by the use of an adequate methodological framework.
The thesis further proposes methods for acquiring microclimatic descriptors that remain useful even in regions with sparse meteorological data, and provides strategies for validating these descriptors in future studies. The most informative analytical scales for these microclimatic data, for both predicting and projecting SDMs, and taking into consideration the spatial differences between training and projection areas, are identified.
The final empirical chapter of the thesis applies the lessons learned from the previous chapters in regard to the optimal implementation of SDMs in islands to analyse the impacts of future climate change on island species using a microclimatic approach. The results of this analysis confirm that endemic species' ranges are disproportionately threatened in most global biodiversity hotspots analysed, and that the current instruments to categorise species risks (e.g., the IUCN Red List) are not able to effectively identify the vulnerability of (island) species to climate change. This research marks a significant advancement in our understanding of island species' responses to climate change at microclimatic scales and offers a crucial baseline for developing more targeted mitigation strategies in different archipelagos around the world.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):

Supervisor(s)	Email	ORCID
Matthews, Thomas	UNSPECIFIED	orcid.org/0000-0002-7624-244X
Sadler, Jon	UNSPECIFIED	orcid.org/0000-0003-0443-4458
Graham, Laura	UNSPECIFIED	orcid.org/0000-0002-3611-7281