Integrating Microclimate, Physiology, Biogeochemistry, and Population Models to Link Climate Change to Demographic Outcomes

Published:

Problem

Climate change impacts species through complex, interacting mechanisms — from environmental conditions to physiological stress, ecosystem processes, and demographic rates.
Understanding these links requires integrating multiple models and data sources into a single, holistic framework that can reveal causality, not just correlation.

Goal: Combine microclimate modelling, mechanistic physiological energetics, biogeochemical pathway modelling, and Bayesian hierarchical population forecasting to mechanistically link climate variability and extremes to recruitment and survival in a climate-vulnerable species.

Approach

  • Microclimate modelling: simulated fine-scale environmental conditions within roosting habitats to capture species-relevant temperature, humidity, and thermal stress.
  • Physiological modelling: quantified energetic and thermal balances to estimate climate-driven physiological stress (e.g. dehydration, overheating) at relevant temporal scales.
  • Biogeochemical modelling: incorporated nutrient cycling and vegetation process models to capture indirect effects on habitat quality and food availability.
  • Population modelling: developed Bayesian hierarchical models linking physiological and biogeochemical predictors to demographic rates (recruitment and survival), explicitly accounting for detection probability in count data.
  • Integrated all components into a unified Bayesian framework implemented in R and JAGS, enabling joint inference and propagation of uncertainty across the entire causal chain.

Stack

  • Holistic modelling integration: combining microclimate, physiology, biogeochemistry, and population dynamics within one framework.
  • Bayesian hierarchical modelling: linking mechanistic covariates to demographic outcomes with full uncertainty propagation.
  • Mechanistic physiological energetics: modelling metabolic and thermal constraints under changing environments.
  • Data workflows: multi-source environmental, physiological, and demographic data cleaning and harmonisation; reproducible pipelines.
  • Implementation: developed in R for data processing, integration, and visualisation; JAGS for Bayesian model specification and inference.

Results

  • Demonstrated causal pathways from climate variability and extremes to population decline through physiological stress on recruitment and survival.
  • Quantified direct and indirect effects, revealing the magnitude of each mechanism and their combined influence on viability.
  • Produced fully integrated forecasts, allowing scenario testing for management interventions.

Impact

  • First application to integrate microclimate, physiology, biogeochemistry, and population dynamics in a unified Bayesian framework for conservation.
  • Provided a mechanistic, evidence-based foundation for targeted conservation planning under climate change.

Links & Resources

Role

  • Designed and implemented the entire multi-component modelling workflow.
  • Developed each model component (microclimate, physiology, biogeochemistry, population) and integrated them into a holistic Bayesian framework.
  • Conducted model fitting, validation, and scenario testing.
  • Authored manuscript, providing the analytical synthesis and causal interpretation.