Now showing items 1-20 of 229

• #### Evaluation of the production potential of Shilaif shale oil formation using analytical and modeling methods

In the early 2000s, the oil and gas industry witnessed a boom of hydrocarbon production from unconventional reservoirs. In 2015, about half of the United States oil production came from unconventional reservoirs (EIA 2017). Unconventional oil reservoirs include, but are not limited to, shale oil reservoirs that are usually the source formations for oil trapped in conventional reservoirs. Due to the huge success of shale oil reservoir development in the US, many other countries around the world, especially countries rich in conventional oil, are currently investigating their unconventional reservoir potential in order to increase their production capabilities to match the world’s ever-increasing demand for oil. Realizing that the importance and contribution of unconventional production to the energy sufficiency in the United States might lead to a similar interest in the UAE in the near future, my research is based on a numerical modeling study using data from a single exploration well in a UAE unconventional shale oil reservoir, namely Shilaif. In addition to modeling, the study includes valuable well test analysis results which helped in assessing reservoir properties. The results of the study indicate that Shilaif has favorable characteristics to be a potential target for shale oil production. However, due to the limited amount of information, data quality, and potential bias in selecting samples, interpreting measurements, and analyzing results, this conclusion should be taken as a preliminary assessment and verified with further data and analysis. One of the most important sources of information to guide future assessments is reliable production and flowback. More core sampling and log-core correlations may also improve the level of confidence in future studies.
• #### Isotopic analyses of helium from wells located in the Four Corners area, southwestern US

Helium, the lightest noble gas, is a valuable resource located on the Colorado Plateau southwestern US. Helium, a proven, useful noble gas, has many applications in modern technology for its chemical, physical, and thermodynamic properties. As of this writing, the price of crude helium is ~40% greater than CH4, rendering the economic grade for direct and secondary extraction at 0.3% helium. The helium systems in the Four Corners area (i.e., the study area) are characterized utilizing the geochemistry of noble gases, hydrocarbons, and non-hydrocarbons (compositional and isotopic), as well as geologic mapping. The geochemistry delineates sources of gases, migration pathways, and potential trapping/sealing mechanisms of the helium system, which is a slight deviation from the petroleum system. Economic helium (>0.3%) is primarily found in Paleozoic intervals structurally trapped on the Four Corners Platform, the edge of the Defiance Uplift, and the edge of the Holbrook Basin. Thirty-one gas samples, isotopically analyzed, are from actively producing Paleozoic formations within five fields: Tocito Dome, Dineh-Bi-Keyah, Ratherford, Pinta Dome, and Navajo Springs. Helium concentrations range from 0.01% to >6.0% and incorporates a spectrum of other gas values associated with relatively similar Paleozoic formations. Noble gases along with hydrocarbon and non-hydrocarbon gas geochemistry are successfully used in genetically fingerprinting gas families. The source of helium is determined to be from the shallow crust, i.e., Precambrian granitic basement. Noted faults and attendant fracture systems serve as primary migration conduits (via fluid flow from advection). Observed gas-water reactions indicate groundwater involvement in the concentration of helium and extreme solubility fractionation (i.e., long secondary migration pathways). By investigating migration pathways, N2 and CO2 are recognized as major, helium carrier gases, whereas CH4 is a helium dilutant. Geologic mapping illustrates dominant structural, stratigraphic, and combination structural-stratigraphic traps. The helium system definition is updated, as well as criteria developed to successfully explore for helium. Proper isotopic and geological analyses can improve helium system models that involve generation, migration, and trapping/sealing mechanisms. Improvements in the understanding of the helium system model are critical for more effective helium exploration. Native Americans Tribes of the Southwest and other economic sectors stand to benefit (economically and socially) from a more enhanced scientific knowledge of the helium system in the study area.
• #### Comprehensive assessment of a hybrid membrane biosystem for sustainable desalination of produced water and frac flowback wastewater, A

Increasing population, agricultural, and industrial growth, compounded by drought in many regions, is exacerbating the stress on existing freshwater resources. Alternative water sources must be identified to alleviate stress and enable sustainable development. The energy sector, particularly the oil and gas (O&G) industry, has the potential to make a substantial impact by reclaiming O&G wastewater and treating it for reuse. Over three million gallons of water can be used to hydraulically fracturing a single well, with up to 40% returning to the surface as wastewater (e.g., produced water (PW)). While this water is typically disposed of via deep-well injection, increasing regulations and environmental concerns are stimulating the development of sustainable and efficient strategies for treatment and reuse. Due to the high and variable total dissolved solids (TDS) in PW, and concentrations of dissolved and free phase organic chemicals and inorganic constituents, robust, multi-barrier treatment approaches are required to achieve reuse standards. Hybrid biological-physical processes could be a promising method for treatment of PW. Biological processes have proven effective at removing organic matter from a variety of waste streams including domestic waste streams, landfill leachate, and oily wastewaters; membranes are most established in desalination of seawater and brackish waters. For sustainable membrane treatment, optimizing water recovery, and reducing membrane fouling, it is critical to implement pretreatment processes that target removal of organic constituents. Several phases of this study have demonstrated high removal of organic matter from PW using biologically active filtration (BAF) with granular activated carbon (GAC) media. This serves as an effective pretreatment for subsequent membrane processes like ultrafiltration (UF) and nanofiltration (NF). Following BAF pretreatment of PW, UF has demonstrated high turbidity removal with minimal membrane fouling, producing high quality permeate. NF then exhibits low fouling propensity, maintaining high ion rejection and permeate flux, producing permeate suitable for advanced reuse applications (e.g., irrigation, streamflow augmentation). Thus, the objective of my dissertation was to assess the hybrid BAF-UF-NF treatment train as a sustainable method for reclamation and reuse of O&G wastewaters. This entailed an overall proof of concept, through evaluation of the treatment train performance (TDS and organic matter removal) with varying qualities of O&G waste streams. Multiple GAC media were compared with addition of nutrients to improve the removal of organic matter by BAF and thus, the overall sustainability of subsequent membrane treatment. The composition of organic matter following BAF, UF, and NF was also characterized to evaluate the removal of specific classes of organic constituents and develop an inexpensive monitoring technique to track organic matter. A long-term investigation of BAF provided additional insight on the feasibility and biological stability with continuous exposure to challenging PW. Evaluation of BAF as a sufficient pretreatment method for desalination of PW was explored through a comprehensive membrane fouling study, providing insight on the potential for reuse of O&G wastewater.
• #### U-Pb and Lu-Hf LA-ICP-MS detrital zircon and structural investigations in the Abitibi subprovince, Canada, with implications for Archean geodynamic processes and deformation behavior along gold-bearing, crustal-scale faults

This work presents investigations based in the south-central Abitibi subprovince of Ontario and Quebec, Canada, which place constraints on a sequence of Neoarchean geodynamic and structural processes, including crustal growth and amalgamation with the rest of the Superior Province, and the formation of orogenic gold deposits along regional deformation zones. Archean geodynamic processes were primarily investigated by multi-isotope U-Pb and Lu-Hf laser ablation – inductively coupled plasma – mass spectrometry (LA-ICP-MS) analysis of detrital zircon grains from successor basins. Three temporally distinct successor basin groups are recognized: the ~2690-2685 Ma Porcupine assemblage, the ~2685-2682 Ma Pontiac subprovince, and the ~2680-2670 Ma Timiskaming assemblage. All samples contain abundant Neoarchean grains (~85-95% of the individual sample populations), while the remaining grains yielded Mesoarchean ages. The Neoarchean grains likely reflect predominately local sources. Since no Mesoarchean rocks occur in the Abitibi and Pontiac subprovinces, and based on comparison with published zircon ages from the Superior Province, Mesoarchean grains are interpreted as derived from an orogenic hinterland to the NNW that developed during regional amalgamation. The older Porcupine assemblage contains ~5% Mesoarchean zircon, while Pontiac subprovince and Timiskaming assemblage samples contain ~18% and ~13%, respectively. This suggests hinterland sources were more prevalent during the late stages of collision, probably as a result of progressive uplift and denudation of the hinterland. The paired Lu-Hf isotopic analyses support this interpretation of provenance and further indicate that the majority of grains (~96%) have compositions consistent with derivation from a modern MORB depleted mantle (mMORB-DM) reservoir. This occurrence of mMORB-DM-like signatures in detrital zircon grains that were sourced from multiple disparate crustal domains suggests that a modern-style depleted mantle reservoir was not only well established, but widely occurring by the Mesoarchean. The observed pattern of lateral accretion of crustal domains that grew from a mMORB-DM-like reservoir, the transport of detritus across terrane boundaries, and the progressive uplift and denudation of a hinterland are most consistent with the operation of plate tectonic processes during Neoarchean construction and amalgamation of the southern Superior Province. During the late stages of accretion, deformation became increasingly localized along regionally extensive, crustal-scale fault zones, including the Larder Lake Cadillac deformation zone (LLCdz) in the Kirkland Lake area of Ontario. New field mapping and map compilation indicates that a series of spaced brittle-ductile deformation zones occur in a broad, >6 km area to the north of the LLCdz. Detailed structural mapping and analysis further indicate that the location of late-stage brittle-ductile deformation, fluid flow, and gold mineralization was likely controlled by early brittle deformation zones characterized by breccia. Therefore, undiscovered gold deposits may be hosted by deformation zones farther from the LLCdz than previously recognized and that fault-related breccia bodies may provide an indicator for nearby prospective structures in both the Kirkland Lake area and in similar structural settings of all ages, worldwide.
• #### Identification of parameters for predicting long-runout landslides in the western United States

No existing research provides an integrated analysis of the key parameters that contribute to long-runout landslides in the Western United States. This study begins the task by assembling a dataset of geological, topographical, and hydrological parameters for landslides from eight study areas. Six measures of mobility were analyzed and two (landslide height drop to runout length ratio, H/L and landslide runout length, L) were selected for further use. Analysis of the correlations of the measured parameters with H/L and L was performed to quantify how well they predict these two mobility measures. The initial slope angle was found to match H/L for small landslides that did not experience a break in slope. Landslides in concave topography, landslides on previously moved material, and landslides in confined topography were found to possess lower H/L values, indicating higher mobility. Finally, landslides occurring on previously moved material and landslides in confined topography were found to possess larger values of L.
• #### Experimental study on the anisotropy of unconventional tight reservoirs: joint ultrasonic and electrical measurements under pressure

Unconventional tight reservoirs have gained importance in global oil and gas production. The fine-grained tight reservoir rocks are often anisotropic as results of their laminar structures in multiple scales of observation. Anisotropic textures of fine-grained rocks have significant influence on pore structures and fluid transport properties. For successful development of unconventional reservoirs, it is critical to precisely characterize evolutions of pore structures and anisotropic flow properties in depleting tight formations. Common geophysical parameters used for characterization of anisotropy in tight reservoirs include wave velocity, attenuation and electrical conductivity. Previous studies have shown that joint investigations using elastic and electrical properties are necessary for assessment of pore structures and permeability anisotropy in tight reservoirs. Conducting these investigations is always challenged by lack of experimental study on anisotropy mechanisms as well as connections between geophysical anisotropy and permeability anisotropy. Furthermore, the study on connection between direction-dependent attenuation mechanism and pore structure is lacking in tight reservoirs. A comprehensive study on anisotropic velocity, attenuation and conductivity responses would greatly benefit characterization of pore structure and flow properties in tight reservoirs. In this thesis I first introduce a new experimental design that provides simultaneous, multi-directional ultrasonic and electrical experiments on cores under pressure conditions. The system is tested and validated using standard references and natural rock samples. Then I investigate individual mechanisms and controls of velocity, attenuation and complex conductivity anisotropy in tight rocks under pressure. Results of ultrasonic measurements show that attenuation and attenuation anisotropy are sensitive to the closure of low aspect ratio pores or microcracks. Attenuation anisotropy in clay and organic rich formations correlates well with the presence of compliant materials (clay and organic matter). Evolution of attenuation anisotropy strongly relates to directional pore connectivity in fine-grained samples. Similarly, electrical textural parameters including formation factor (F) and tortuosity (T) tensor also show correlation with pore deformation and pore connectivity change in tight samples. Tortuosity is a very sensitive parameter for preferential pore alignment and directional pore deformation in tight reservoirs. Finally, this thesis provides results of joint velocity, attenuation and conductivity measurements on sandstone and fine-grained tight rocks under pressure. Results on sandstone imply correlation between rock physics parameters and pore deformation as well as permeability change. For tight reservoir rocks, permeability anisotropy could have completely opposite trends in chalks and shales depending on texture and pore size distribution. Among all elastic, anelastic and electrical anisotropy parameters, imaginary conductivity shows the strongest relationship with preferential pore alignment and directional pore connectivity. While the combination of all three anisotropies provides a comprehensive understanding of textural anisotropy, conductivity anisotropy is the most sensitive parameter for permeability anisotropy assessment and stress monitoring in unconventional tight reservoirs.
• #### Spatio-temporal assessment of groundwater resources in the Denver Basin Aquifer System

Groundwater is an important resource in the Unites States and provides about 40% of the country’s public water supply. Withdrawals have dramatically increased in many aquifers, leading to groundwater depletion and questions about future sustainability. In Colorado, the long-term sustainability of the Denver Basin Aquifer System is considered by some as questionable and insufficient to support future demands. Groundwater depletion has been widely documented over the past several decades as groundwater withdrawals have increased and competition for water further stresses supplies. Groundwater monitoring is fundamental to understanding system dynamics, trends in storage, and the long-term sustainability of an aquifer. However, groundwater level data are typically spatially and temporally sparse relative to the data density desired for aquifer-scale analysis. The problems with missing temporal data from water wells in particular has not been addressed in much detail, yet can cause important misinterpretation with regard to groundwater sustainability. This research aims to mitigate some of the problems with current approaches to analyzing water well data by incorporating a new method of spatial-temporal analysis, with particular emphasis on addressing missing temporal data. In addition, we evaluate the ability of the Gravity Recovery and Climate Experiment (GRACE) satellites to improve the temporal sustainability analysis. The methodology is first illustrated using a case study in the Arapahoe Aquifer and is then expanded to all aquifers of the Denver Basin Aquifer System. Remote sensing is utilized from GRACE to provide another perspective on determining groundwater storage changes. Results from this dissertation provide a framework for monitoring and management of groundwater resources along the Colorado Front Range as well as other water-stressed regions of the western U.S.
• #### Investigation of low oxygen HSLA steel weld metal

Hot-wire gas tungsten arc welding is a process gaining more popularity in industry today due to increased deposition rate, from the resistively heated filler metal introduced by an external wire feed system. A modified version of this system, which incorporates an oscillation mechanism to modify the frequency with which the filler metal enters the molten weld pool was utilized in this work and the effect of varying process parameters on microstructure, weld bead morphology, and inclusion size and distribution are characterized. Hot-wire amperage, wire feed speed, frequency of wire oscillation and heat input were varied on seven wire consumables to determine their effects on acicular ferrite. Optimal microstructural development in high strength low alloy steel weldments is mainly dependent on acicular ferrite, which nucleates on oxide inclusions. However, with the gas tungsten arc process, it is difficult to manipulate weld pool oxygen content to achieve the needed level for optimal acicular ferrite formation. The oxide inclusion population is investigated to determine the morphology, size and spatial distribution and whether they served as nucleation sites for ferrite formation. The microstructure was primarily affected by heat input and composition, with increasing heat input reducing the total acicular ferrite volume percent. These effects directly correlated to inclusion population, with three distinct distributions demonstrated between inclusion radii in the range of 0.1 to 0.75 micron. Composition effects were tied to both an increase in Molybdenum, 0.1 wt. pct. to 0.25 wt. pct., a Titanium peak at 70 ppm and Oxygen content. The oscillation affected weld bead morphology, improved weld wetting but with little effect on final microstructure. Two experimental wire compositions were identified to move forward for additional mechanical testing.
• #### Essays in energy and environmental economics

The chapters in this dissertation focus on environmental economics but vary in their contributions to current environmental economics literature. The second chapter shows that coal stockpiles accumulated through contractual obligations during periods of low natural gas prices restricts the responsiveness of coal-fired plants when relative coal-to-gas prices change. It implies that the social marginal benefit associated with higher price of coal can be smaller than what we would expect in the short run. In the other two chapters, I revisit moving costs, a central concept in non-market valuation. Specifically, I emphasize the role that moving costs play in household migration and their impact on households' exposure to local air pollution. In chapter 3, I find that low-income households face higher moving costs. I run simulation models to show that the heterogeneous moving costs can explain why a disproportionate number of low-income people live near undesirable land uses. Chapter 4 offers a complementary argument to Chapter 3. I use a natural break in moving costs among Californian homeowners to test if lower moving costs cause households to move to cleaner neighborhoods in terms of air pollution. I do not find such evidence from the public-use census data set, but the low geographic resolution of the data might have driven the result.
• #### Controls on deposition, lithologic variability, and reservoir heterogeneity of prolific western interior shelf sandstone reservoirs: Tocito and El Vado sandstones, San Juan Basin, NM

The Tocito and El Vado Sandstones have proven to be highly prolific hydrocarbon reservoirs in the largest domestic onshore conventional gas basin in the U.S., the San Juan Basin (SJB). Application of modern drilling technology and favorable petroleum commodity prices have resulted in increased exploration interest in the SJB. Despite 50 years of development and extensive research, our understanding of the distribution and variability of productive reservoir facies and the factors influencing reservoir heterogeneity is unclear. The Tocito Sandstone is a locally deposited, coarse-grained, glauconite rich sandstone with a unique depositional style involving multiple depositional sequences and complex erosional contacts due to focused tidal influence in local paleogeographic lows. The El Vado Sandstone is a low porosity, low permeability regressive-transgressive, storm wave-influenced shelf sand that produces in vertical wells from natural fractures and extends deep into the SJB. Despite being commonly encountered in wells as stacked reservoir intervals, the Tocito Sandstone and El Vado Sandstone are lithologically unique, stratigraphically separate and require different technologies to exploit. Ample opportunity exists for new exploration in both, especially in areas previously un-explored using unconventional techniques. Through extensive regional well log correlations, core analysis and thin section petrography, this study characterizes the controls on deposition of the Tocito and El Vado Sandstones, the regional variability of the nature of these sandstones, the vertical relationships of the Tocito and El Vado Sandstones and underlying units and the influencing factors on reservoir properties and exploration success.
• #### Uncertainty quantification in seismic imaging

To make informed decisions, one has to consider all available knowledge about the assessed problem. An important part of the decision-making process is understanding uncertainties and how they influence the outcome. In seismic exploration, many decisions are based on interpretations of seismic images, which are affected by multiple sources of uncertainty. Thus, image uncertainty quantification is an important, albeit challenging task. In this thesis, I focus on two uncertainty sources that affect seismic imaging: data uncertainty and velocity uncertainty. I quantify the seismic data uncertainty using theoretical analysis applied to two field experiments with repeated shots. My analysis reveals that amplitude distributions for each data sample as a function of time and position are not Gaussian and that the uncertainty of a seismic event is proportional to its mean amplitude. I also find that seismic events excited by the source are highly repeatable, but small changes of the source position impact the amplitude response, highlighting the importance of geometry repeatability for the lapse studies. Velocity uncertainty also has a large impact on image uncertainty, as it affects reflector positioning and the focusing of seismic events. By examining two subsalt imaging scenarios with geological uncertainty caused by the salt body physical properties, I demonstrate that image uncertainty, expressed as a function of the image amplitude or as a function of the reflector location, is the largest under the salt: the image amplitude distributions are two times broader under the salt than away from it. The confidence index maps are a useful tool to convey the information about image amplitude uncertainty to an interpreter, while the location uncertainty reveals uncertain directions and is affected by acquisition geometry. The main challenges facing uncertainty quantification in seismic imaging include integration of different sources of uncertainty and reducing the computational cost of the analysis. My analysis leads to recommendations about possible approaches towards these challenges, with emphasis on using sparsity to reduce the dimensionality of the problem.

• #### Redox cycles with doped calcium manganites for high-temperature thermochemical energy storage in concentrating solar power

Redox cycles with reducible perovskite oxides of the form ABO$_3$ can provide thermochemical energy storage (TCES) with higher energy density and storage temperatures than molten-salt systems for large-scale energy storage in concentrating solar power (CSP). Perovskites from earth abundant cations are desirable for cost-effective solutions, but such materials must demonstrate appropriate thermodynamics for high specific TCES and favorable kinetics for heat-driven reduction and exothermic re-oxidation. This dissertation explores the thermodynamics and kinetics of doped CaMnO$_{3-\delta}$ particles for TCES redox cycles where particles are heated and reduced in N$_2$ ($P_{\text{O2}} \approx 10^{-4}$ bar) to high temperatures (700 to $1000^{\circ}$C) in a solid-particle solar receiver. Chemical and sensible energy stored in the reduced perovskite particles is released as needed to a supercritical CO$_2$ power cycle via re-oxidation and cooling of the material. Thermodynamics of Ca$_{1-x}$Sr$_x$MnO$_{3-\delta}$ ($x=0.05$ and $0.1$) and CaCr$_y$Mn$_{1-y}$O$_{3-\delta}$ ($y=0.05$ and $0.1$) are characterized through thermogravimetric analysis and calorimetry. Results indicate Ca$_{1-x}$Sr$_x$MnO$_{3-\delta}$ compositions can store over 200 kJ kg$^{-1}$ more specific energy storage compared to inert particulate TES media for $T \ge 900^\circ$C; the specific energy storage potential of Ca$_{0.9}$Sr$_{0.1}$MnO$_{3-\delta}$ at $T=900^\circ$C and $P_\text{O2}=10^{-4}$ bar is 706 kJ kg$^{-1}$. Challenges are expected achieving these high values of energy storage in a transport-limited receiver with low residence time for CSP. Redox kinetics are explored in a packed bed reactor with rapid heating capabilities. Results in isothermal tests show that oxidation is significantly faster than reduction. Modeling of packed bed experiments indicate that reduction at $T \ge 800^\circ$C is limited by build-up of oxygen in the gas phase and equilibrium thermodynamics between the solid and gas phases. Long-term redox cycling tests, which simulate a nominal TCES cycle, demonstrate excellent chemical stability for all materials. A standard deviation of 1.9\% on the extent of reduction over 1000 cycles was observed for Ca$_{0.9}$Sr$_{0.1}$MnO$_{3-\delta}$. Modeling efforts of the packed bed experiments allow for characterization of redox kinetics, to be implemented in computational models for system component design. One of the most promising compositions, Ca$_{0.9}$Sr$_{0.1}$MnO$_{3-\delta}$, is implemented in a 1-D receiver model to explore designs and operating conditions for perovskite-based energy storage systems.
• #### Tempering response of mixed martensitic/bainitic microstructures in quench and tempered plate steels

High strength heavy gauge plate steels develop different cooling rates throughout the thickness during quenching. Several grades do not fully harden to martensite, resulting in mixed microstructures of martensite and bainite, which are subsequently tempered to achieve required properties. Therefore, comprehension of the tempering of bainitic and mixed microstructures is essential to help optimize tempering conditions for plate steels. The tempering response of martensitic, bainitic, and mixed microstructures was examined with a focus on alloying effects induced by molybdenum (Mo), vanadium (V), chromium (Cr) and silicon (Si). Hardenability characterization was conducted through Jominy end quench testing and construction of continuous cooling transformation (CCT) diagrams. Increased alloying additions, in particular Mo and Cr, improved hardenability and hence increased the allowable thickness of fully hardened plate steels. Characterization was performed using scanning electron microscopy (SEM), dilatometry, Mössbauer spectroscopy, X-ray diffraction (XRD), electrical resistivity, Vickers micro-hardness, tensile and Charpy V-notch testing. A dilatometric analysis of non-isothermal tempering was proposed, which was able to evaluate tempering stages I through III, as well as characterize secondary hardening and Mn segregation to cementite. Application of a rule of mixtures approach using results for fully martensitic and bainitic tempered microstructures was able to predict the hardness of the mixed conditions in most cases. The presence of bainite in the initial microstructure yielded an improved tempering resistance, which was related to the lower driving force and number of nucleation sites for cementite precipitation. Secondary hardening was affected by isothermal bainitic transformation temperature and Cr and Si additions. The secondary hardening peak shifted to higher temperatures for a steel isothermally transformed at higher temperatures. Peak secondary hardening was postponed and less pronounced with Cr addition in martensitic and bainitic steels, respectively, which is believed associated with the segregation of Cr to cementite, retarding its dissolution. Silicon is responsible for a change in the secondary hardening mechanism in bainitic steels by reducing the extent of cementite precipitation, permitting solute carbon to be readily available. Thus, peak secondary hardening shifted to lower tempering temperatures in martensitic steels, and greater softening rates were observed in the bainitic conditions.
• #### Representation learning for long-term collaborative autonomy

Autonomy has attracted a lot of research attention over the past few decades, since it is the key capability of all autonomous systems, including unmanned aerial vehicles (UAV), unmanned ground vehicles (UGV), unmanned surface vehicles (USV), humanoid robots, etc. Those fully or partially autonomous systems have been transforming the way people work, live, and communicate nowadays, e.g. automated AC systems in the building, robot arms manufacturing cars in the factory, etc. On the other hand, robots or intelligent agents usually do not work alone, such as assistance robots, coaching robots, self-driving cars, etc. They need to observe, learn from, and reflect with human beings. When robots enable to interact and collaborate with humans autonomously, we call it collaborative autonomy. Collaborative autonomy is a very challenging problem, which requires robots to have both great perception and decision making capabilities. It becomes even more challenging when this collaborative autonomy can be continuously performed in a long-term period, since there would be strong appearance variations of the environment, such as changes of illumination, weather, and vegetation conditions across months or even seasons. Humans can easily identify the same object and place in different times of the day, months, and seasons. However, this critical long-term perception capability is very challenging for real-world robots though it is the key to enable long-term autonomy. This research investigates the perception problems for long-term collaborative autonomy. In this dissertation, several representation learning approaches are introduced to improve the real-time perception performance of robots in the long-term period. Firstly, I introduce a 3D human skeletal representation learning approach to enable real-time robot awareness of human behaviors, which is invariant to viewpoint, human body scale and motion speed. Then, multiple representation learning approaches are presented for the long-term place recognition problem, which enables the life-long relocalization of robots with a single camera. Finally, we demonstrate that the learned representation using the approaches proposed in this dissertation can be integrated in the online robotic decision making system and enables the long-term collaborative autonomy capability.
• #### Biochemical and chemical controls on sedimentation, sequence stratigraphy, and diagenesis, in the Phosphoria rock complex (Permian), Rocky Mountain region, USA

Biochemical sedimentation, near-surface diagenesis, stable isotopes, and porosity vary systematically stratigraphically and regionally in the Phosphoria Rock Complex (PRC), Rocky Mountain region, USA, in response to dynamic paleo-environmental conditions spanning the Middle Permian. Environmental, biochemical, and isotopic trends mimic diagnostic trends of the End-Permian Mass Extinction (EPME), occurring through Kungurian-Wordian time (~274Ma to 265Ma), ~13MY before the EPME and ~5MY before the end-Guadalupian crisis. These indicate PRC trends are of a similar genesis to the EPME, and EPME dynamics were driven by locally modified global processes spanning the Middle to Late Permian. Biochemical, isotopic, and environmental trends are heterogeneous across the PRC, a second-order (~9MY) cycle, and the third-order (2-5 MY) Franson (latest Kungurian - Wordian) and Ervay (Wordian) cycles. During transgressions, influx of cool, acidic, low-oxygen, nutrient-rich waters warmed and interacted with hot, oxygenated marine and evaporitic waters. Flourishing sapropelic algal and anaerobic microbial communities resulted in phosphorites and sulfidic-OM-rich mudrocks seaward of calcitic biota and micritic carbonates, redbeds, evaporites and microbialites. Values of δ18O and δ13C in carbonates and silica are depleted in distal settings due to microbial decay of OM and increase landwards due to evaporative fractionation. Values of δ18O in carbonate fluorapatite are depleted in distal environments, increase landwards and towards maximum transgression, indicating warming. Porosity in transgressions is low due to near-surface cementation and recrystallization, as well as compaction and infill of pore space by secondary OM (bitumen) in OM-rich mudrocks. S-rich OM catalyzed early secondary-OM generation and inhibited OM-hosted porosity generation through burial. During highstands warm, oxygenated, alkaline marine waters dominated in and became increasingly hot, shelf-confined, and evaporitic. With limited nutrient influx, and input of eolian-sourced silica, resulted in widespread spiculites and calcitic-biota carbonates at maximum transgression and in distal highstands. Increasingly restricted and evaporitic conditions through highstands resulted in dolomitized bioturbated muds and sandstones, aragonitic molluscs, ooids, and peritidal microbialites. Values of δ18O and δ13C became increasingly enriched throughout highstands in marine carbonates and widespread authigenic silica. This and moganite-bearing chalcedony suggest evaporitic reflux drove silicification and dolomitization. Porosity is most abundant in dolomites deposited in restricted, evaporitic highstand conditions.