京都大学 / 新制・課程博士 / 博士(工学) / 甲第23479号 / 工博第4891号 / 新制||工||1764(附属図書館) / 京都大学大学院工学研究科社会基盤工学専攻 / (主査)教授 立川 康人, 准教授 市川 温, 教授 角 哲也 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

The Little Creek watershed was assessed to identify changes in event-based suspended sediment export and determine the factors contributing to sediment production the first year following the Lockheed Fire in 2009. The amount and volume of near-stream sediment sources were found to decrease, while an increase in hillslope sediment production was documented. High intensity, short duration rainfall (up to 87 mm/hr for 10 minute duration) initiated extensive rilling and minor channel-derived debris torrents originating from the upper south facing slopes. Rainfall simulations, hillslope erosion plots, and soil infiltration tests indicated that fire produced soil water repellency, the lack of ground cover, steep slopes, and high soil burn severity were the most influential factors contributing to hillslope erosion. Contrary to results reported in other western U.S. studies, regression analyses determined that the effect of fire significantly decreased suspended sediment concentrations with higher flows at North Fork and Upper North Fork monitoring stations. The effect of the fire did not produce increases in stormflow volumes and event sediment load, likely due to the fact near-stream sediment contribution was minimal and the majority of hillslope-derived sediment sources were not hydrologically connected. This study provides valuable information for landowners and land managers to understand how a coastal redwood dominated watershed responds to wildfire and prepare post-fire mitigation efforts following future wildfires.

Les particules en suspension transportées dans les réseaux hydrographiques résultent des processus de rivière et des apports depuis les versants. Nous avons étudié dans cette thèse la dynamique des tailles de particules le long du continuum versant-rivière afin d'apporter des éléments de réponse à la réflexion aujourd'hui menée sur les distances de transport et sur le concept de connectivité sédimentaire. Des observations de terrain sont menées à l'exutoire d'un bassin versant de tête (~20 km²). Elles mettent en évidence une corrélation positive entre débit liquide et taille des particules. L'établissement et la mise en oeuvre d'un protocole de mesure original montre que les particules sont agrégées. A cette échelle, les apports des versants semblent importants pour expliquer les variations de taille des particules. Des expériences de laboratoire utilisant un canal annulaire sont menées et indiquent qu'une partie de ces variations peut être attribuée à la désagrégation ou à la floculation des particules dans l'écoulement. Des variations de taille notables sont dues au type de sol. Elles sont moindres à la fin des évènements schématiques simulés en canal, suggérant que l'écoulement prend une part prépondérante pour expliquer les variations de taille des particules. Cet effet du type de sol a motivé l'étude des processus de versant, et en particulier ceux de la pluie. Des expériences de simulations de pluie menées en laboratoire (~1 m²) sur deux sols révèlent que l'augmentation de l'énergie cinétique de la pluie a tendance à générer des agrégats plus fins. Une paramétrisation du détachement par la pluie par fraction granulométrique est développée sur la base de ces expériences et implémentée au sein de deux modèles numériques d'érosion hydrique à base physique. Les simulations numériques confirment que cette tendance a des impacts sur les exports à l'échelle du versant. Enfin, des variations de granulométrie en lien avec l'énergie cinétique de la pluie sont perceptibles lors d'observations de terrain à l'échelle du versant (~ 100 m²), confirmant l'importance d'une description correcte du forçage pluviométrique. / The suspended particles of catchment networks are dependent on both river and hillslope erosion processes. During this thesis, the particle size dynamics was studied along this continuum in order to improve the understanding of particle delivery from hillslopes to the outlets of headwater catchments. Field measurements were conducted at the headwater catchment scale (~20 km²). The discharge displayed a positive correlation with the particle size. An original measurement protocol has been set up and it demonstrated that particles were mostly aggregated. The inputs from hillslopes were possibly involved in some of the variations of the measured particle size. Laboratory experiments carried out using an annular flume demonstrated that a part of these variations could be explained by disaggregation or flocculation within the flow. Important variations due to the soil type were observed. However, they were less pronounced in the falling limbs of the schematic flood events, suggesting that flow conditions progressively became more important than the soil signature. The latter encouraged the analysis of hillslope processes, among which a special attention was given to the rainfall effects. Rainfall simulation experiments (~1 m²) demonstrated for two soils that an increase in the rainfall kinetic energy resulted in smallest aggregates detached from the soil matrix. The importance of this mecanism at the hillslope scale (~ 100 m²) with regard to runoff selectivity was demonstrated developing a size-dependent detachment parametrisation included in two physically based numerical models. Finally, the effects of the rainfall kinetic energy on the particle size were observed during field measurements made at the plot scale as well, underlining the need to adequatly describe the rainfall forcing field at this scale.

<p>Soil organic carbon (OC) is one of the most important terrestrial carbon pools and plays a major role in climate regulation, water quality, provisional services, and numerous other ecosystem functions. The conversion of natural vegetation and the supporting soil to intensively managed agricultural systems put soil at risk for loss due to erosion and enhanced microbial degradation with loss rates increased by orders of magnitude above the pre-managed system. The process has negatively impacted agricultural productivity on hillslopes by diminishing soil health, as well as the quality of stream water and coastal aquatic environments, and it is an important but as of yet poorly quantified factor in the region’s terrestrial C budgets. There have been substantial debates on the role of erosional and depositional processes on the landscape as a control on exchange of C between the land surface and the atmosphere. A central aspect of the debate stems from the limited data regarding the fate of soil erosion-induced transport of OC through stages of detachment and splash, transport and redistribution, deposition and burial. The overarching purpose of this thesis is to evaluate how dynamic patterns of soil OC erosion due to intensive agricultural management influences soil aggregate strength, the chemical nature of mobilized organic particles, and connectivity and sourcing between hillslope and streams. Using both simulated and natural, short-term, event-based erosive rainfall processes, with a multiproxy geochemical approach, we attempt to develop a comprehensive understanding of how upland watershed mechanistic controls soil movement and associated chemical alterations to the material exported through dissected segments from hillslope to the fluvial network. </p> <p>Our results demonstrate that erosive processes on hillslope connects between terrestrial sources to receiving potential deposition settings, actively ‘filter’ soil aggregates and particles and associated OC at each erosional stage (i.e., detachment and transport downhill/downstream), with distinct geochemistry in low relief and poorly drained agricultural systems, like the CCW. Complex interactions among tillage intensity, tillage practice-induced, oriented surface roughness, and storm-induced hydrological connectivity, that potentially impact the fate of these transported OC upon decomposition, deposition and burial, and have important implications for predicting landscape level heterogeneity in surface and buried soil chemistry upon mobilization and burial, as well as the dynamics of sourcing and transformation of material exported to inland water systems.</p>