Prioritizing removals of a highly cryptic invasive species using machine learning and demographic weighting.

Authors

Alexander S Romer, Melissa A Miller, Sergio A Balaguera-Reina, Michelle C Bassis, Brandon Welty, Michael Kirkland, Amy Peters, Edward F Metzger, Jenna Cole, LeRoy Rodgers, Frank J Mazzotti

Published in

Ecological applications : a publication of the Ecological Society of America. Volume 36. Issue 6. Pages e70317.

Abstract

Invasive species present substantial challenges to biodiversity conservation, particularly when low detectability and demographic heterogeneity obscure the population-level consequences of management actions. Invasive Burmese pythons (Python bivittatus) in South Florida exemplify this challenge, as metrics based on raw removals or body size alone may not reflect demographic impacts. To address this, we introduce the Weighted Removal Index (WRI), a demographic weighting framework that prioritizes removals based on expected contributions to survival and reproduction. Using survey data from the South Florida Water Management District's Python Elimination Program, we quantified operational and environmental conditions associated with higher survey WRI, defined as the summed WRI of all pythons captured on a single survey, and evaluated three forecasting approaches: Generalized Additive Models (GAMs), Boosted Regression Trees (BRT), and Neural Networks (NN). Survey WRI was highest under waning lunar phases, when surveys were assisted, and under warmer and more humid conditions. Additionally, survey WRI was elevated on warm days early in the wet season and cool dry-season days. In held-out test evaluation, the NN achieved the highest balanced accuracy for predicting survey success (0.703), and the lowest error for predicting survey WRI (root mean squared error [RMSE] = 5.50). Contractor performance was highly heterogeneous, with contractor-specific effects positively associated with mean survey WRI (R² = 0.589) and capture rate (R² = 0.596). Notably, model-based predictions of survey success fell within the upper tail of the contractor capture-rate distribution, indicating that near-term forecasting can match the performance of the highest performing contractors. Together, demographic weighting and near-term forecasting provide a practical decision-support framework for guiding effort toward high-impact removals in cryptic invader management.

PMID:
42767653
Bibliographic data and abstract were imported from PubMed on 22 Sep 2026.

Read full publication at:
Please sign in to see all details.

Stats

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.