arXiv subjects

Michael R. Kearney

Publications and source records attributed to Michael R. Kearney.

3 recordsLinked to original sources

arXiv2026

Spatially Aware Deep Learning for Microclimate Prediction from High-Resolution Geospatial Imagery

Microclimate models are essential for linking climate to ecological processes, yet most physically based frameworks estimate temperature independently for each spatial unit and rely on simplified representations of lateral heat exchange. As a result, the spatial scales over which surrounding environmental conditions influence local microclimates remain poorly quantified. Here, we show how remote sensing can help quantify the contribution of spatial context to microclimate temperature predictions. Building on convolutional neural network principles, we designed a task-specific deep neural network and trained a series of models in which the spatial extent of input data was systematically varied. Drone-derived spatial layers and meteorological data were used to predict ground temperature at a focal location, allowing direct assessment of how prediction accuracy changes with increasing spatial context. Our results show that incorporating spatially adjacent information substantially improves prediction accuracy, with diminishing returns beyond spatial extents of approximately 5-7 m. This characteristic scale indicates that ground temperatures are influenced not only by local surface properties, but also by horizontal heat transfer and radiative interactions operating across neighboring microhabitats. The magnitude of spatial effects varied systematically with time of day, microhabitat type, and local environmental characteristics, highlighting context-dependent spatial coupling in microclimate formation. By treating deep learning as a diagnostic tool rather than solely a predictive one, our approach provides a general and transferable method for quantifying spatial dependencies in microclimate models and informing the development of hybrid mechanistic-data-driven approaches that explicitly account for spatial interactions while retaining physical interpretability.

cs.LG↗

arXiv2024

The untapped power of a general theory of organismal metabolism

What makes living things special is how they manage matter, energy, and entropy. A general theory of organismal metabolism should therefore be quantified in these three currencies while capturing the unique way they flow between individuals and their environments. We argue that such a theory has quietly arrived -- 'Dynamic Energy Budget' (DEB) theory -- which conceptualises organisms as a series of macrochemical reactions that use energy to transform food into structured biomass and bioproducts while producing entropy. We show that such conceptualisation is deeply rooted in thermodynamic principles and that, with the help of a small set of biological assumptions, it underpins the emergence of fundamental ecophysiological phenomena, most notably the three-quarter power scaling of metabolism. Building on the subcellular nature of the theory, we unveil the eco-evolutionary relevance of coarse-graining biomass into qualitatively distinct, stoichiometricially fixed pools with implicitly regulated dynamics based on surface area-volume relations. We also show how generalised enzymes called 'synthesising units' and an information-based state variable called 'maturity' capture transitions between ecological and physiological metabolic interactions, and thereby transitions between unicellular and multicellular metabolic organisation. Formal theoretical frameworks make the constraints imposed by the laws of nature explicit, which in turn leads to better research hypotheses and avoids errors in reasoning. DEB theory uniquely applies thermodynamic formalism to organismal metabolism, linking biological processes across different scales through the transformation of matter and energy, the production of entropy, and the exchange of information. We propose ways in which the theory can inform trans-disciplinary efforts at the frontiers of the life sciences.

q-bio.PE↗

arXiv2022

Mechanistic forecasts of species responses to climate change: the promise of biophysical ecology

A challenge in global change biology is to predict how species will respond to future environmental change and to manage these responses. To make such predictions and management actions robust to novel futures, we need to accurately characterize how organisms experience their environments and the biological mechanisms by which they respond. All organisms are thermodynamically connected to their environments through the exchange of heat and water at fine spatial and temporal scales and this exchange can be captured with biophysical models. Although mechanistic models based on biophysical ecology have a long history of development and application, their use in global change biology remains limited despite their enormous promise and increasingly accessible software. We contend that greater understanding and training in the theory and methods of biophysical models is vital to expand their application. Our review shows how biophysical models can be implemented to understand and predict climate change impacts on species' behavior, phenology, survival, distribution, and abundance. We illustrate the types of outputs that can be generated, and the data inputs required for different implementations. Examples range from simple calculations of body temperature to more complex analyses of species' distribution limits based on projected energy and water balances, accounting for behavior and phenology. We outline challenges that currently limit the widespread application of biophysical models. We discuss progress and future developments that could allow these models to be applied to many species across large spatial extents and timeframes. We highlight how biophysical models are uniquely suited to solve global change biology problems that involve predicting and interpreting responses to environmental variability and extremes, multiple or shifting constraints, and novel abiotic or biotic environments.

q-bio.PE↗