Plant temperature in a Mediterranean woodland savanna: Measurements and models

Abstract

Vegetation temperature plays a central role in ecosystem science at scales from cells to biomes: enzymatic reactions underpinning plant growth and rates of carbon sequestration are temperature-dependent, leaf temperature is a key determinant of the rate at which plants lose water, and temperature strongly mediates exchanges of energy between the biosphere and the atmosphere. However, despite well-developed and long-standing biophysical theory, the challenges of measuring vegetation temperature at ecologically-meaningful time and space scales have largely stymied the study of spatially-explicit thermal patterns and the evaluation of surface energy balance in ecosystem models. In this thesis, I use diverse tools to explore patterns of vegetation temperature in a Mediterranean woodland savanna and to assess and advance thermal measurements and models.

In Chapter 1, I extend the theory of near-earth thermal remote sensing to open-canopy ecosystems, and I present the first long-term time series of thermal images in a woodland savanna. I characterize thermal imaging challenges that are particularly acute in the open-canopy context. Specifically, emissivity and background radiation data required for image calibration are variable across the scene, mixed pixels can be especially misleading because of divergent component temperatures, and targets of interest have very different pixel dimensions associated with their different distances from the camera. Results from this chapter underpin the collection of accurate thermal imaging data in open-canopy ecosystems.

In Chapter 2, I quantify vertical temperature gradients within tree canopies and between over- vs. under-story Mediterranean savanna plants. I describe a surprising pattern of intra-canopy temperature: mid-day savanna tree canopy tops are cooler than canopy bottoms, associated with the under-story grass reaching daytime temperatures exceeding over-story temperatures by up to 16.3 degrees Celsius. I then determine the relationship between satellite thermal data and vertically-resolved, ground-based measurements of plant temperatures, to assess the ability of nadir-looking instruments to measure temperature in this ecosystem.

In Chapter 3, I evaluate a cohort-based terrestrial biosphere model with near-earth, airborne, and satellite thermal remote sensing data. I find that, while modeled surface temperatures respond appropriately to seasonal and diurnal meteorological forcing, they are consistently several degrees Celsius too warm in the dry season. This divergence between model and data is likely associated with the model's lack of an energetically-resolved layer of plant litter. Diagnosis of model/data correspondence provides avenues for model development and vital context as we apply terrestrial biosphere models to predict ecosystem function in a changing world.

In sum, this thesis integrates novel field measurements, remote sensing images, eddy covariance, an ecosystem model, and literature to develop our understanding of and abilities to measure and model plant temperature in a Mediterranean woodland savanna.