A big data and machine learning approach for monitoring the condition of ecosystems

This is a Preprint and has not been peer reviewed. This is version 3 of this Preprint.

Add a Comment

You must log in to post a comment.


There are no comments or no comments have been made public for this article.


Download Preprint


Miguel Equihua , Octavio Pérez-Maqueo, Julián Equihua, Pedro Maeda, Michael Schmidt, Melanie Kolb, Nashieli Garcia Alaniz, Mariana Munguía-Carrara, Sergio Villela, Oliver López,, Everardo Robredo, Santiago Martínez


Ecosystems are highly valuable as a source of goods and services and as a heritage for future generations. Knowing their condition is extremely important for all management and conservation activities and public policies. Until now, the evaluation of ecosystem condition has been unsatisfactory and thus lacks practical implementation for most countries. We propose that ecosystem integrity is a useful concept that can be used to evaluate ecosystem condition through data science and machine learning. Based on a three tier (contextual, instrumental and hidden) model and a Bayesian network approach, we used field and remote sensing data to estimate the integrity of terrestrial ecosystems per 250 m in Mexico.




Applied Statistics, Biodiversity, Bioinformatics, Earth Sciences, Ecology and Evolutionary Biology, Environmental Sciences, Forest Biology, Forest Sciences, Life Sciences, Other Ecology and Evolutionary Biology, Physical Sciences and Mathematics, Statistical Methodology, Statistical Models, Terrestrial and Aquatic Ecology


Ecosystem Integrity, data science, machine learning, Bayesian Network, terrestrial ecosystems, Environmental Accounting, SEEA-United Nations


Published: 2024-01-16 09:08

Older Versions

CC BY Attribution 4.0 International

Additional Metadata


Data and Code Availability Statement: