Spatiotemporal Patterns in Soil Carbon Distribution, Persistence, and CO2 Respiration in a Pacific Northwest Montane Forest

Given their considerable ability to store and stabilize carbon (C), soils are a critical resource to maintain in the face of the accelerated effects of climate change on natural systems. Pacific Northwest montane forests are hotspots of above- and belowground C storage globally, yet the combined effects of extended seasonal drought, longer and more extreme fire seasons, and warming temperatures are already disrupting C cycling in these systems and stimulating release of soil C to the atmosphere. Accelerated release of CO2 may intensify C-temperature feedbacks, fueling further global changes. The processes that influence organic matter accumulation and C stabilization and destabilization in forest soils are still not fully understood, and can be especially hard to tease apart in areas with complex interactions among spatial and temporal drivers of soil C. Quantifying landscape-scale (km) and finer grain (m) resolution estimates of soil C, soil C cycling rates, and drivers of soil C stabilization and destabilization can help inform ecosystem models that feed into land management decisions. The H.J. Andrews Experimental Forest (HJA) is a 6400-ha Long Term Ecological Research (LTER) site in Oregon’s western Cascade Mountains with complex terrain, varied vegetation assemblages, steep slopes, and a substantial gradient in elevation. HJA hosts the Detrital Input and Removal Treatment (DIRT) experiment, which has manipulated organic matter input rates of needle litter, woody debris, and root-derived C for over two decades. The first chapter of this dissertation examines the effects of sustained additions or removals of detritus on soil respiration to address questions about the longevity of the soil organic matter priming effect. While adding a more labile C source in the form of needle litter resulted in slightly increased release of soil CO2 beyond the amount expected from litter additions, there were compensatory gains in soil C relative to control treatments. I provide evidence that soil organic matter priming is a short-term phenomenon and that there are more likely seasonal changes in moisture availability that are driving changes in plant and microbe soil C allocation. Surprising diurnal trends in soil respiration illustrate the tightly regulated relationship between tree stomatal conductance, midday vapor pressure deficit, and root-derived soil respiration. Additionally, root and rhizosphere respiration contributed the most to total respiration, while above- and belowground decomposition of organic matter contributed less. The second chapter of this dissertation expands in scope to a spatial analysis of soil carbon distributions across mid- and high- elevations of HJA. This research addresses questions about the interplay between topographic and vegetative drivers of persistent and labile soil C pools in complex terrain. Ratios of C to nitrogen (N) tended to be greater in valley versus ridge sites and were much greater in forest versus meadow sites, which may be an artifact of the N-limitations in lower elevations where nutrients are cycling more quickly – in contrast to higher elevations where decomposition is slowed and vegetative growth is less resource-limited. In chapter three, I further expanded the study area to the entirety of HJA and investigated the large- and small-scale drivers of total soil C and of soil C fractions, in addition to soil N. I found that, unsurprisingly, soil depth was the most significant predictor of soil C, but that important environmental controls included elevation (as a proxy for temperature and moisture regimes) and proximity to the nearest stream. I was surprised to find that aboveground biomass and landscape position were less important to prediction of soil C relative to climate thresholds. Using a combination of my field data and machine learning techniques, I produced maps of soil C, N, and mineral-associated and particulate organic matter C distributions across HJA. I compared my mapped soil C products with publicly available soil datasets and found wide variation in predicted soil C across different datasets that can be explained in part by their coarse resolution and interpolation across too few field samples. The insights gained from the studies in this dissertation point to the importance of matching the spatial and temporal scale of sampling to the scale of ecological processes, a critical step in producing higher resolution estimates of soil C across complex landscapes.