Open grassy environments in Africa have been key landscapes for the development and evolution of humans and our hominid ancestors for millions of years. These environments have not been static, however, as global climate changes have strongly shaped their nature and location over time. In the modern, at least 80 million people living in Sub-Saharan Africa rely on agriculture and pastoralism within the grasslands and savannas of the Sahel region alone for food security. Devastating droughts and associated famine in the region over the past several decades have highlighted this region’s potential vulnerability to future climate change. Wildfires play a unique and critical role in maintaining Africa’s grasslands and savannas, especially in the Sahel region. Emissions from these fires have additional ramifications for the Earth’s radiative balance and global cycles of carbon and nutrients. As populations in Africa rise over the coming century from ~1.3 billion to 4 billion people by 2100, increasing demand for food, rising temperatures, and highly uncertain changes in rainfall and wildfire patterns are poised to put the people and ecosystems of this region in jeopardy. In the face of potentially novel environmental conditions resulting from anthropogenic climate change, this research aims to better understand the long-term interconnections of climate, ecology, and human presence in Northwest Africa and how these linkages may vary under broad shifts in climate. Accurate projections of future climate and ecosystem change not only require the mechanistic understanding of climate forcings and climate-ecosystem interactions that can be gleaned from modern relationships, but also information about how these interactions may vary as a function of changes in the background climate state itself (on centennial to million year timescales). Highly spatially-resolved satellite measurements relevant for asking such questions only extend back a few decades and thus only provide a limited perspective on whether or not modern climate-ecosystem interactions are stationary through time. This thesis is focused on developing and applying paleoclimate reconstruction techniques to generate new records of hydroclimate, ecosystem structure, and fire activity in Northwest Africa over a broad set of time scales. These new records are used to assess the governing controls of climate variability and evaluate the evolution of climate-ecosystem interactions across a diversity of background climate states. We seek specifically to (1) improve our understanding of the natural climate forcings that dictate changes in Northwest African monsoon rainfall, (2) evaluate how changes in rainfall and other climate parameters––namely atmospheric CO₂ concentrations––together affect ecosystem distributions and compositions in Northwest Africa, (3) ground-truth the use of increasingly popular molecular proxies of fires applied to marine sediment archives, (4) assess the relative environmental and human controls on fires in Northwest African savannas over time, and (5) develop interpretive frameworks for understanding multiproxy records of environmental changes in Northwest Africa to draw conclusions about how climate-ecosystem interactions may have evolved over time. To address these goals, this dissertation is broken down into four chapters. The first two chapters focus on the orbital-scale to multimillion year forcings in the climate system that control the strength and tempo of the Northwest African monsoon and how these changes impact the distributions and compositions of ecosystems in the region over time. In both Chapters 1 and 2, we develop new reconstructions of hydroclimate using the hydrogen isotopic composition of plant-waxes and extraterrestrial 3He normalized dust fluxes from marine sediment core MD03-2705 taken off the coast of Mauritania along the Northwest African margin. We further reconstruct ecosystem change using the carbon isotopic composition of plant-waxes. Chapter 1 is centered on the late-Pleistocene while Chapter 2 takes a wider perspective and explores long-term trends in Northwest African hydroclimate and vegetation structure from the Pliocene to the late-Pleistocene. In the second half of this work, the focus is narrowed to center the role of fire in Northwest African savannas and how the nature of burning in this region has changed since the last glacial maximum. In Chapter 3, we use atmospheric back trajectory modeling and a transect of marine core sediments taken aboard the research vessels Vema and Conrad that spans the southern European to southern west African margin to test if molecular biomarkers of vegetation (plant-wax n-alkanes and pentacyclic triterpene methyl ethers) and fires (pyrosugars and polycyclic aromatic hydrocarbons) preserved in marine sediment archives capture modern distributions of ecosystems and biomass burning on the landscape. In Chapter 4, we generate new records of the fire history of Northwest Africa from the last glacial maximum to the late-Holocene from marine sediment core OCE 437-7 GC68. We compare the relative influences of changes in rainfall, ecosystem structure and human activities on fire across the most recent deglaciation and ‘Green Sahara’ period. From these new records, we are able to draw several conclusions. In Chapter 1 we find that the location, timing, and intensity of northwest African monsoon rainfall is controlled by low-latitude insolation gradients and that while increases in precipitation are associated with the expansion of grasslands into desert landscapes, changes in pCO2 predominantly drive the C3/C4 composition of savanna ecosystems. In Chapter 2 we observe that low latitude insolation gradients best explain both the tempo and amplitude of orbital scale variations in Northwest African rainfall over the last five million years, however strengthening sea surface temperature gradients in the Atlantic Ocean during the mid-Pleistocene likely led to a precipitous and sustained decline in monsoon strength ~900 thousand years ago independent of any change to orbital insolation forcing. Furthermore, changes in the relationship between rainfall and vegetation in Northwest Africa can be used to track changes in the northward extent of ecosystems, augmenting previous pollen-based reconstructions, which together show shifts in ecosystem distributions over the Plio-Pleistocene likely related to changes in ecosystem disturbances and climate-vegetation interactions. In Chapter 3 we show that by accounting for the effects of long-range transport of biomarkers, good agreement is found between ecosystem composition and biomass burning patterns on the African continent and the distribution of terrestrial plant and fire biomarkers in marine core top sediments. This provides strong justification for applying molecular indicators of fires to the paleorecord. In Chapter 4, we show that rainfall is the dominant natural control on the amount of biomass burned in Northwest African savannas, but increased human presence and land-use change during the mid- to late-Holocene likely fundamentally changed the fire regime of Northwest Africa to this day.

In this study, I addressed how a large-scale wildfire affected a prairie arthropod community in southern Alberta, Canada. First, I looked at the general effects of disturbance on the arthropod community. Second, I addressed how processes such as competition and secondary succession may have affected diversity in this arthropod community. Third, I determined how the arthropod community trophic structure was regulated. Results showed that the effect of disturbance on arthropods varies greatly by taxa. Factors, such as site, year, distance from disturbance edge, as well as the disturbance itself, were important in determing the abudance, biomass, richness, and diversity of the arthropod community. There were strong year-to-year differences hat exceeded the disturbance effects. Results showed that the processess behind the intermediate disturbance hypothesis, succession and competition are not detectable in this arthropod community. This arthropod community was likely regulated in a bottom-up manner, in which herbivores ultimately control the abundance of predators and parasitoids. / x, 97 leaves : ill. ; 28 cm.

This study provides an analysis of watershed response to climate change and forest fire impacts, to better understand the hydrologic budget and inform water management decisions for present and future needs. The study site is 2,365 km2, located in the upper Umatilla River Basin (URB) in northeastern Oregon. The Precipitation Runoff Modeling System, a distributed-parameter, physical-process watershed model, was used in this study. Model calibration yielded a Nash Sutcliffe Model Efficiency of 0.73 for both calibration (1995-2010) and validation (2010-2014) of daily streamflow. Ten Global Climate Models using Coupled Model Intercomparison Project Phase 5 experiments with Representative Concentration Pathways 4.5 and 8.5 (RCP), were used to observe hydrologic regime shifts in the 2020s, 2050s, and 2080s. Mean center timing of flow occurs earlier in the year in both pre- and post-fire conditions, where there are increased winter flows and decreased summer flows throughout the 21st century. Change in temperature and percent change in precipitation is more variable in the summer than winter increasing over time, with a slight decrease in winter precipitation in the 2080s in RCP 8.5. Temperature increases 1.6°C in RCP 4.5 and 3.3°C in RCP 8.5 by the end of the 21st century. The ratio of Snow Water Equivalent to Precipitation decreases 96% in the 2080s in RCP 8.5 before forest cover reduction, and decreases 90-99% after forest cover reduction. Potential basin recharge and the base-flow index are both sustained throughout the 21st century with slight declines before forest cover reduction, with an increase in basin recharge and increase in base-flows in the 2080s after fire-burns. However, the simulated sustained base-flows and area-weighted basin recharge in this study, do not take into account the complex geologic structure of the Columbia River Basalt Group (CRBG). A more robust characterization and simulation of URB aquifer recharge would involve coupling the PRMS model with a groundwater model in a future study. Although groundwater recharge in the CRBG in the URB is not well understood, the long-term decline of groundwater storage presents a serious environmental challenge for the Confederated Tribes of the Umatilla Indian Reservation and communities in the URB.

Forested watersheds in western North America are subject to significant change from natural and anthropogenic disturbance, including wildfire. Forest canopy changes have subsequent impacts on sub-canopy snow processes. A simple, process-based point energy balance model was developed to quantify differences in energy balance characteristics between a burned and a healthy forest stand. Potential model uncertainties were identified using sensitivity analyses. Simulated snowmelt accurately recreated measured snowmelt, providing confidence in the model’s ability to simulate energy balance processes in subcanopy environments where wind redistribution and sublimation are not major drivers of the local snowmelt energy balance. In the burned stand, sub-canopy snow accumulation was greater but melted more rapidly than in the healthy stand. The removal of forest canopy resulted in more energy available for snowmelt, including higher short-wave and lower long-wave radiation, and increased turbulent fluxes. Burned stands should be considered a separate land cover type in larger scale watershed models. / xii, 129 leaves : ill,, map ; 29 cm