The Earth’s climate is influenced by complex interactions of physical, chemical, and biological processes that link terrestrial ecosystems and the atmosphere. One of these interactions involves the use of light in photosynthesis, which allows plants to remove CO2 from the atmosphere and slow the unprecedented rate of climate change the Earth is experiencing. However, modeling future climate remains challenging, in part because of limited knowledge of mechanisms controlling the effects of light on gross ecosystem CO2 uptake (conceptually, photosynthetic activity integrated across all leaves in a plant canopy). Unlike previous studies, this dissertation uses data from atmospheric science, ecosystem ecology, and plant physiology to provide evidence for mechanistic links between physical, biophysical, and ecological controls on the effects of light on processes tied to gross ecosystem CO2 uptake—specifically, ecosystem gross primary production (GPP) and leaf photosynthesis. First, this dissertation empirically demonstrates that the dominant effect of clouds is to reduce total light above canopies. However, optically thin clouds increase scattered, diffuse light, which canopies use more efficiently than they use direct light. This offsets reductions in total light and results in no net change in GPP under thin clouds, while GPP decreases under optically thick clouds because both diffuse and direct light decrease. Second, ground-based measurements indicate that the rate of increase in GPP with diffuse light changes throughout the day. The magnitude of increase depends on how canopies interact with the angle of incoming light to biophysically alter the distribution of light within canopies and thus, the proportions of leaves contributing to GPP. Third, the distribution of species and light within one forest canopy leads to differences in some of the rate-limiting biochemical reactions in leaf photosynthesis. These field-based data indicate which assumptions representing canopies in Earth system models may not have support in situ, and could be contributing to errors in model estimates of future climate. Overall, this dissertation identifies mechanisms through which clouds and plant canopy structure alter land-atmosphere CO2 fluxes and subsequently, Earth’s climate. It also provides an important interdisciplinary framework for testing assumptions about the feedbacks that living organisms form with their environment.