Dissertations, Theses, and Capstone Projects

Date of Degree


Document Type


Degree Name





Robert P. Anderson

Committee Members

Ana C. Carnaval

Phillip Staniczenko

Diego Alvarado-Serrano

Lázaro Guevara

Subject Categories

Biodiversity | Other Ecology and Evolutionary Biology | Zoology


Climate change, Extrapolation, Maxent, Range shift, Species distribution models, Time series


Ecological niche models (ENMs) are commonly used to estimate the potential geographic distributions of species and have various uses, such as predicting invasive species' potential distribution, assessing vulnerability to climate change, or estimating paleodistributions. However, ENM developments have primarily focused on solving spatially related issues, underemphasizing the temporal dimensionality of occurrence and environmental data collected over the last century. One limitation of ENMs is the common practice of using recent environmental averages that fail to capture the gradual environmental change over the time span of the occurrence records and can hide trends in environmental variables, a critical piece of information that could help understand the impact of ongoing climate change. Moreover, the traditional ENM framework has been criticized for its limited use of the timing of species occurrences, which is typically associated with a single average value of the environmental condition over a long period of time (e.g., 30-year climate baselines) even if the records corresponded to decades earlier. Another challenge occurs when transferring models to different time periods with dissimilar environmental conditions from those in which the model was trained. Despite available recommendations for transferring to novel conditions, tools that facilitate decision-making focused on characteristics of the model chosen for transfer are lacking. To help fill these gaps and using two montane small non-volant mammals in Mexico, i) I introduce a new method for detecting changes in environmental suitability by using a time series of recent data, ii) examine the effectiveness of temporal matching between occurrences and environmental conditions, and iii) present examples and visualization tools to help in the decision-making process and implementation when transferring to novel environmental conditions.

In this dissertation, I present three chapters that collectively expand the role of the temporal dimension in ecological niche modeling. In the first chapter, I focus on predicting potential changes in the distribution limits of the Mexican small-eared shrew (Cryptotis mexicanus). Instead of comparing model predictions between two time periods (present vs. a future scenario), I used a time series of environmental data over the past four decades to identify temporal trends of environmental suitability. The findings indicate that changes in environmental suitability do not align with the simple poleward or upslope shifts expected for a montane species. Instead, variation in regional precipitation, rather than temperature, is the primary factor that corresponds with changes in suitability affecting the distribution limits. The approach used here could be a valuable supplement for forecasting potential geographic range changes across different time periods. In the second chapter, I temporally match occurrences of C. mexicanus with the environmental conditions experienced prior to the date of observation (i.e., one, five, and ten years) and compared resulting models against those from a standard 30-year average to evaluate their performance and ecological plausibility. The results showed that the ten-year temporal resolution performed equally to or better than the standard approach (based on withheld omission rate). Temporal matching could therefore improve model performance and geographic predictions, even for species with low mobility. Future studies should focus on selecting an optimal temporal resolution that considers population responses to climate change. In the third chapter, I use the Black-eared mouse (Peromyscus melanotis) as a case study for transferring models to novel (“non-analog”) conditions, also known as environmental extrapolation. I develop visualization tools to help in the decision-making process for extrapolating models when transferring to novel environments. This study reaffirms the importance of inspecting response curves and simultaneously assessing the quantity of pixels under novel conditions to make appropriate choices regarding extrapolation methods. The recommendations in this study offer additional guidance for obtaining more reliable geographic predictions in novel conditions, which is becoming increasingly crucial in a rapidly changing world. Overall, this dissertation contributes to the methodological development of incorporating the temporal dimension in ecological niche modeling through empirical examples and code that can be used broadly by biogeographers and spatial ecologists.

This work is embargoed and will be available for download on Tuesday, September 30, 2025

Graduate Center users:
To read this work, log in to your GC ILL account and place a thesis request.

Non-GC Users:
See the GC’s lending policies to learn more.