IMPROVED UNDERSTANDING AND MODELING OF GROSS PRIMARY PRODUCTION AND ECOSYSTEM RESPIRATION OF TERRESTRIAL ECOSYSTEMS: DIURNAL, SEASONAL, AND INTERANNUAL DYNAMICS

Abstract

Terrestrial ecosystems play a critical role in the global carbon cycle. At the ecosystem scale, green vegetation assimilates atmospheric carbon dioxide (CO2) through photosynthesis, a process commonly quantified as gross primary production (GPP). Simultaneously, terrestrial ecosystems release CO2 back to the atmosphere through ecosystem respiration (ER), which encompasses both autotrophic and heterotrophic components. The difference between GPP and ER defines the net ecosystem exchange (NEE) of CO2, which represents the net carbon flux between the terrestrial biosphere and the atmosphere. As a key indicator of ecosystem carbon balance, accurate estimation of NEE can help determine whether an ecosystem functions as a carbon sink or source, with direct implications for global carbon budget accounting and climate change mitigation. Satellite observations offer the advantages of global coverage and high temporal revisit frequency, making satellite-based models a promising approach for regional to global estimation of GPP, ER, and NEE. Meanwhile, the growing accumulation of eddy covariance (EC) observations has created new opportunities to explore the responses of carbon fluxes to environmental drivers and to advance model development. In this dissertation, by integrating in situ EC observations from the global FLUXNET network with satellite remote sensing data, I sought to: (1) investigate the light absorption by chlorophyll; (2) characterize the temperature response of ER; (3) develop a data-driven Vegetation Photosynthesis and Ecosystem Respiration Model (VPERM) capable of simultaneously estimating GPP, ER, and NEE at half-hourly temporal resolution; and (4) assess ecosystem carbon sink and source dynamics at decadal scales. Chapter 2 introduced the Vegetation Photosynthesis Model v3.0 (VPM v3.0), which incorporated two major modifications: (1) the introduction of a site-specific apparent optimum air temperature parameter for GPP, and (2) a revised water stress function to better capture environmental limitations on vegetation productivity. VPM v3.0 was applied at five long-term EC flux tower sites located in grassland and savanna ecosystems across the United States to generate 8-day GPP estimates for the period 2000–2021. Chapter 3 investigated the light absorption by chlorophyll. Chapter 4 explored the temperature response of ER and identified a site-specific apparent optimum temperature for ER. Chapter 5 combined these advanced understandings to develop VPERM and applied it to hundreds of EC tower sites across diverse biomes globally. Chapter 6 applied VPERM, driven by ERA5 climate reanalysis data from 2000 to 2024, to examine carbon sink and source characteristics and their interannual variability at undisturbed terrestrial ecosystems. Most of these undisturbed ecosystem sites, including old-growth forests, functioned as carbon sinks during 2000–2024. Substantial interannual variation in annual NEE was also observed, primarily driven by large climate variations and asymmetric responses of GPP and ER to hydrothermal stresses. These results demonstrate the importance of conserving undisturbed terrestrial ecosystems and carrying significant implications for nature-based climate solutions and carbon–climate feedback.