ADVANCING FOREST CARBON MONITORING AT HIGH SPATIAL AND TEMPORAL RESOLUTION FOR REGIONAL APPLICATION

Abstract

Forests play a critical role in the global carbon cycle, yet accurate regional carbon monitoring remains challenging due to fine-scale spatial and temporal heterogeneity in carbon stocks, fluxes, and tree cover. Existing observational approaches are often limited in their ability to capture this heterogeneity simultaneously across space and time, creating gaps in our understanding of regional carbon budgets. This dissertation advances high-resolution regional forest carbon monitoring by developing and applying high spatial and temporal approaches across three complementary studies in Maryland, USA. First, I developed the Maryland Forest Carbon Flux Digital Twin (MFCF-DT) at 90-meter spatial and hourly temporal resolution to evaluate the representativeness of existing flux sampling networks, finding that both the SERC eddy covariance tower and CARAFE airborne campaigns had limited statewide representativeness. To improve the sampling coverage, I proposed a new CARAFE flight track, which was subsequently flown in summer 2025. Second, I quantified trees outside forests (TOF) at 30-meter resolution from 2011 to 2023 using integrated remote sensing and ecosystem modeling, revealing that TOF contribute approximately 23% of statewide tree cover and 15% of aboveground carbon stocks, representing a substantial but previously untracked component of Maryland's carbon budget. Finally, I applied a Siamese convolutional neural network to bi-temporal NAIP aerial imagery, demonstrating the feasibility of early detection of newly planted trees in support of Maryland's Five Million Trees Initiative. While technical barriers remain, the framework offers a scalable complement to field-based monitoring and is transferable to similar tree planting programs in other regions. Together, these studies demonstrate that fine-scale spatial and temporal data reveal critical gaps in regional carbon monitoring and offer practical pathways to more complete and accurate forest carbon accounting.