Erica Julissa Cortez

Artificial Intelligence and Biodiversity Conservation in the Borderlands

Along the United States–Mexico (US–MX) border, steel walls disrupt natural ecological
connections by blocking wildlife movement (e.g., migration) and limiting interactions with native
plants. Quail calls that once carried across the desert meet a barrier, breaking interactions
needed for survival. The border cuts across the habitats of more than 1,000 native species,
including federally endangered animals. This isolates habitats and blocks migration routes
needed for reproduction and long-term population stability.

This project takes an interdisciplinary approach by integrating large historical datasets, citizen
science observations gathered by the public, and recorded bird vocal datasets to examine how
desert flora along the US–Mexico border supports quail species that indicate desert health. This
research will also investigate plant associations such as velvet mesquite and evaluate how the
separation of habitats limits quail movements in this underfunded region.

AI will analyze quail occurrence, bird vocals, and plant–habitat data from GBIF and SEINet across
the US–MX borderlands. Machine learning (ML) will identify patterns between plants and quail
species. A convolutional neural network (CNN) will detect and classify quail calls by visualizing
patterns in the wavelength of sound, converting them to images to train the model. Interactive
maps will highlight high activity areas to identify conservation hotspots, and a user-friendly app
will allow users to upload recordings to identify quail species using the trained CNN model. This
approach helps build a framework for large-scale conservation, further engages the public in
scientific discovery and demonstrates how AI technology can support ecological research.

Developing an interactive map and user-friendly app will allow residents, conservation groups, and 
researchers to access and visualize biodiversity, promoting active participation in monitoring and 
protecting desert habitats. There is an ethical and cultural obligation to conserve the borderlands’ 
natural ecosystems for present and future generations to experience them.

Personally, Erica will gain hands-on experience analyzing biodiversity data, modeling species–
habitat relationships and applying bird call monitoring. Publishing the code and trained AI model
on Zenodo, along with a short PLos or Royal Society publishing paper, will strengthen Erica's
academic profile. This experience will prepare her for a future in STEM research that requires
expertise in coding, data science, and community collaboration.