Forest Biomass and Soil Carbon Stock Changes in the Delaware River Basin

Forests play an important role in the global carbon cycle. Quantifying forest biomass and soil carbon stocks and their change over time and space is important to understand forest dynamics and their feedbacks with climate change. This dissertation investigates the forest biomass and soil carbon stocks and their controlling factors in the Delaware River Basin (DRB) using a combination of field measurements and modeling. In 2001-2003, 77 forest plots in three research sites were established and their biomass was measured. In 2012-2014, 61 of these plots were revisited and forest biomass was re-measured using the same protocols. Two soil sampling methods, the Forest Inventory Analysis standard soil core method and the quantitative soil pit method, were also used to collect soil samples. Based on the results of field measurements, a process-based ecosystem model (PnET-CN) was parameterized and used to simulate the spatial distribution of forest carbon pool and fluxes in the three sites. We found that the mean biomass carbon stock in the three sites was 166.5 Mg C ha-1 and had increased by 2.35 Mg C ha-1 yr-1 and was thus a carbon sink over the past decade. The soil carbon stock to 40 cm depth was 76.6 Mg C ha-1. The accuracy of the soil core sampling method was questioned because in the surface mineral soil layer, lower bulk density, lower coarse fragment content and greater carbon concentration were measured using the core method compared to the pit method. By parameterizing the wood turnover rate, maximum photosynthesis rate and disturbance year based on field measurements, the performance of the PnET-CN model was improved in capturing the spatial variation of forest carbon dynamics. The modified model was also used in experimental scenarios, demonstrating 39% of forest carbon sequestered over the past decade could be attributed to the combined effects of elevated CO2 and nitrogen deposition. Large uncertainties in forest carbon stocks at regional scales are associated with the spatial heterogeneity of the forest. A long-term forest monitoring system combined with modelling can greatly reduce the uncertainties and increase the accuracy of our estimates of forest carbon stocks