Comparing the performance of UAV-derived statistical models with established mechanistic models to measure microclimatic conditions in open and closed canopy environments

Climate change is reshaping ecosystems, yet regional climate data often overlook the fine-scale temperature patterns that species actually experience. These local variations, or microclimates, are influenced by vegetation, topography, and land cover, and can shield organisms from extreme conditions. This project aims to improve how we model and map microclimates using high-resolution drone imagery. We will develop statistical models that transform thermal infrared (TIR) and landscape data such as canopy cover, land cover, and landscape heterogeneity into detailed maps of near-surface air temperature. Using hourly drone imagery and ground air temperature records, we will compare these statistical models to established mechanistic models that simulate temperature from physical processes. By testing both open and forested habitats, we will identify how vegetation affects model accuracy and determine the best approach for capturing real microclimate variation. This collaboration combines the intern’s unique expertise in drone-based thermal mapping with Dr. Lenoir’s experience in forest microclimate modelling, strengthening research capacity at both institutions. The project will advance the use of remote sensing in ecological monitoring, contribute new methods for predicting how species respond to climate change locally, and create valuable international partnerships that help develop microclimate science in Canada and introduce new methods abroad.

Faculty Supervisor:

Jeremy Kerr

Student:

Partner:

Université de Picardie Jules Verne

Discipline:

Life Sciences

Sector:

Education

University:

University of Ottawa

Program: