Climate Change and Local Host Availability Drive the Northern Range Boundary in the Rapid Expansion of a Specialist Insect Herbivore, Papilio cresphontes

J Keaton Wilson,Dominique Berteaux,Rebecca A Hutchinson,Maxim Larrivée,Kent P Mcfarland,Kathleen L Prudic,Nicolas Casajus,Jeremy T Kerr

doi:10.3389/fevo.2021.579230

Abstract

Species distributions, abundance, and interactions have always been influenced by human activity and are currently experiencing rapid change. Biodiversity benchmark surveys traditionally require intense human labor inputs to find, identify, and record organisms limiting the rate and impact of scientific enquiry and discovery. Recent emergence and advancement of monitoring technologies have improved biodiversity data collection to a scale and scope previously unimaginable. Community science web platforms, smartphone applications, and technology assisted identification have expedited the speed and enhanced the volume of observational data all while providing open access to these data worldwide. How to integrate and leverage the data into valuable information on how species are changing in space and time requires new best practices in computational and analytical approaches. Here we integrate data from three community science repositories to explore how a specialist herbivore distribution changes in relation to host plant distributions and other environmental factors. We generate a series of temporally explicit species distribution models to generate range predictions for a specialist insect herbivore (Papilio cresphontes) and three predominant host-plant species. We find that this insect species has experienced rapid northern range expansion, likely due to a combination of the range of its larval host plants and climate changes in winter. This case study shows rapid data collection through large scale community science endeavors can be leveraged through thoughtful data integration and transparent analytic pipelines to inform how environmental change impacts where species are and their interactions for a more cost effective method of biodiversity benchmarking.

Highlights

Biodiversity benchmarking is fundamental to both basic and applied ecological research offering insights into the biological processes shaping species and their interactions
We focused on eastern North America where Papilio cresphontes has been reported to be expanding rapidly (Finkbeiner et al, 2011; Breed et al, 2012) and data are readily available for both P. cresphontes and larval host plants, (Zanthoxylum americanum, Zanthoxylum clava-herculis and Ptelea trifoliata)
We model the distributions of the butterfly’s naturally occurring larval host plants, which, when combined with analysis of P. cresphontes range, result in different conclusions for the future distribution of this butterfly than if we had relied on abiotic variables alone (Figures 2, 3)

Summary

Introduction

Biodiversity benchmarking is fundamental to both basic and applied ecological research offering insights into the biological processes shaping species and their interactions. Observational web platforms and smartphone applications, automated camera arrays, and machine learning-assisted identifications have changed how biodiversity data is collected, processed, and verified (e.g., Sullivan et al, 2009; Prudic et al, 2017) challenges remain (Bonney et al, 2009). These technologies have expedited the rate of understanding and changed the research focus to exciting new areas where an informatics toolkit is a necessity (Feng et al, 2020). One new aspect of benchmarking biodiversity is to evaluate where species are and which species they co-occur with, or species distributions and their changing interactions (e.g., Bueno de Mesquita et al, 2016; Palacio and Girini, 2018)

Methods

Results

Discussion

Conclusion