A Multi-Scale Assessment of Solar-Induced Chlorophyll Fluorescence and Its Relation to Northern Hemisphere Forest Productivity

Abstract

Photosynthesis, or gross primary productivity (GPP), plays a critical role in the global carbon cycle, since it is the sole pathway for carbon fixation by the biosphere. Quantifying GPP across multiple spatial scales is needed to improve our understanding of current and future behavior of biosphere-atmosphere carbon exchange and subsequent feedbacks on the climate system. Remote sensing represents one method to observe vegetation properties and processes, and solar-induced chlorophyll fluorescence (SIF), a light signal originating from leaves, has been shown to be proportional to GPP on diurnal and seasonal timescales. Recently, new techniques to retrieve SIF from satellite observations have provided an unprecedented opportunity to study GPP on a global scale. The relationship between SIF and GPP, however, is subject to significant uncertainty as it is influenced by a number of ecosystem traits (e.g. plant species, canopy structure, leaf age). In this dissertation, I evaluate SIF signals and their relation to GPP over Northern Hemisphere forest ecosystems.

First, I compare climate-driven variations in satellite-based SIF to both longstanding satellite vegetation indices derived from reflected sunlight and tower-based estimates of GPP. Even when aggregated regionally, interannual variability (IAV) of SIF is found to be subject to low signal-to-noise performance, particularly during summer. However, through a statistical analysis, I show that increases in springtime temperature driven by warmer temperatures are offset by drier, less productive conditions later in the growing season. Summer productivity, however, is more strongly correlated with moisture than with temperature, suggesting that moisture exerts a greater influence on growing season-integrated signals. While these results demonstrate that satellite observations can be used to reveal meaningful carbon-climate interactions, they also show that currently available satellite observations of SIF do not allow for robust studies of IAV at scales comparable to surface-based observations.

To investigate how SIF signals are related to ecosystem function at a local scale, I built and deployed a PhotoSpec spectrometer system to the AmeriFlux tower at the University of Michigan Biological Station (US-UMB) above a temperate deciduous forest. These observations show a strong correlation between SIF and GPP at diurnal and seasonal timescales, but SIF is more closely tied to solar radiation and exhibits a delayed response to water stress-induced losses in summer GPP. This decoupling during drought highlights the challenges in using SIF to detect changes in summertime productivity. However, an increased ratio between red and far-red SIF during drought indicates that the combination of SIF at multiple wavelengths may improve the detection of water stress. Lastly, I explore diurnal and directional aspects of the SIF signal. Observations of SIF are sensitive to sun-sensor geometry, with smaller incident angles (between solar and viewing angles) leading to stronger signals. However, afternoon SIF is typically lower than morning values at equivalent light levels due to ecosystem downregulation, which obfuscates angular dependencies in the afternoon. While satellite observations typically rely on a clear-sky sunlight proxy to scale instantaneous observations of SIF to daily values, these results demonstrate the need to account for sounding geometry and diurnal hysteresis in SIF signals in order to advance the interpretation of satellite observations. Overall, my results provide a multiscale assessment of SIF over Northern Hemisphere forests and emphasize that careful attention must be given to the spatial and temporal scales at which SIF can be used to make inferences about GPP.