Repository logo

Browsing by Author "Ham, Jay M., committee member"

Now showing 1 - 19 of 19
  • Thumbnail Image
    ItemOpen Access
    A social-ecological approach to managing agricultural ammonia emissions and nitrogen deposition in Rocky Mountain National Park
    (Colorado State University. Libraries, 2017) Piña, Aaron Joshua, author; Denning, A. Scott, advisor; Ojima, Dennis S., advisor; Schumacher, Russ S., committee member; Baron, Jill S., committee member; Ham, Jay M., committee member
    Atmospheric nitrogen (N) deposition is harmful to nutrient-limited mountain ecosystems. Annual wet deposition of total inorganic N in Rocky Mountain National Park (RMNP) is dominated by ammonium, which primarily comes from agricultural sources. The most wet N deposition events between 1980 and 2015 occurred during summer months. The confluence of summertime mountain meteorology and the location of pollution sources are a perfect combination that leads to high values of wet N deposition in RMNP. In Chapter 2, we tested the importance of convection as a N transport mechanism in addition to large-scale east winds, typically associated with the summertime mountain-valley circulation on the eastern plains of Colorado. We characterized the meteorological transport by using the Weather Research and Forecasting model at 4/3-km horizontal resolution. We used passive tracers as a simplified representation of emissions from a single agricultural source in eastern Colorado during three summer precipitation events where wet N deposition values in RMNP were among the highest recorded in all summers between 1980 and 2015. In all three cases, anticyclones in north-central United States and monsoonal flow associated with the North American Monsoon brought together the necessary conditions for deep convection over RMNP. Output from our simulations suggested large-scale winds were responsible for slow and steady transport whereas convection was a rapid and intermittent form of transport. This chapter showed two scales of transport had an additive effect that led to high deposition of N in RMNP during the afternoon/evening hours of three case studies. Chapter 3 discusses the development of a pilot early warning system (PEWS) for agricultural operators to voluntarily and temporarily minimize emissions of NH3 during periods of upslope winds. The PEWS was created using trajectory analyses driven by outputs from an ensemble of numerical weather forecasts together with the climatological expertise of human forecasters. In this study, we discuss the methods for the PEWS and offer a preliminary analyses of 21 months of the PEWS based on deposition data from two sites in RMNP as wells as voluntary responses from agriculture managers and producers after warnings were issued. Results from this study showed that the PEWS accurately predicted 5 of 7 high N deposition weeks at the lower-elevation observation site, but only 3 of 8 high N deposition weeks at the higher-elevation observation site. With the higher-elevation site receiving pollution from sources both west and east of the Continental Divide, sources west of the Continental Divide would need to be included in the PEWS to capture all of the sources leading to deposition at the higher-elevation site. Sixty agricultural producers and managers from 39 of Colorado's agricultural operations volunteered for the PEWS, and a two-way line of communication between the producers and the scientists was formed. An average of 21 voluntary responses (s.d. 4.9) per warning occurred, with over 75% of the PEWS participants altering their practices after an alert. Solving a broad and complex social-ecological problem requires both a technological approach, such as the PEWS, and collaboration and trust from all participants, including agricultural producers, university researchers, and environmental agencies. Chapter 4 applies a systems approach that explores the actors involved in a complex social-ecological problem that deals with the competing interests of an unadulterated environment and the contribution towards feeding the global population. Agricultural operations in northeastern Colorado are among the densest in the world. The demand of a growing global population has put pressure on the agricultural community to provide large quantities of food in a short amount of time. The cost for higher yields means more water, nutrients, and energy, and the result is environmental degradation in the forms of atmospheric and water pollution. The problem becomes more complex when we mix bottom-up and top-down management approaches. That is, agricultural producers are asked to work together with state and federal agencies on reducing emissions from their operations. A pilot early warning system employed in Colorado since 2014 helped bring together the actors to work towards the common goal of reducing nitrogen deposition in Rocky Mountain National Park. Our goal in this chapter was to organize the problem using a conceptual, social-ecological framework. The case studies and pilot early warning system from Chapters 1 and 2 document starting points for how institutional decisions can incorporate agricultural stakeholders in a mix of bottom-up and top-down management approaches under current and future climatic conditions.
  • Thumbnail Image
    ItemOpen Access
    An investigation of ammonia and inorganic particulate matter in California during the CalNex campaign
    (Colorado State University. Libraries, 2012) Schiferl, Luke D., author; Heald, Colette L., advisor; Collett, Jeffrey L., committee member; Ham, Jay M., committee member
    Over the last century, the rise of industrial agriculture has greatly increased the emission of ammonia (NH3) from livestock waste and synthetic crop fertilizers to the atmosphere. Ammonium (NH4+) aerosol, which can be formed through the neutralization of atmospheric acids by NH3, is a key component of particulate matter (PM) in the atmosphere. PM causes negative human health effects and reduces visibility, and transport and deposition of excess NH3 can cause environmental degradation. Airborne observations of gas precursors and inorganic aerosol taken during the CalNex campaign in May and June 2010 are used in this study to investigate the role of NH3 in PM formation in California and test the representation of the key processes relevant to this chemical system in the GEOS-Chem chemical transport model. Evaluation of the 0.5° x 0.667° horizontal resolution nested model with observations shows a large underestimation (5.4 ppb median bias in the lowest 1 km) of NH3 in the Central Valley. This NH3 underestimation is lower in the area surrounding Los Angeles (LA), only 1.4 ppb. Sulfur dioxide (SO2) is also underestimated in both regions, while nitric acid (HNO3) shows little bias. Near-surface simulated inorganic PM is under-predicted by 1.28 µg sm-3 in the LA region and over-predicted in the Valley by 0.27 µg sm-3. Investigation of model sensitivity to the processes of gas-particle partitioning, wet deposition, dry deposition and emissions reveal that emissions have the largest potential for correction of model deficiencies. Increases to anthropogenic livestock NH3 emissions by a factor of 5 and anthropogenic SO2 emissions in the Valley by factors from 3 - 10 eliminates the bias in the simulation of gases in both regions and PM near LA, where under-prediction of nitrate (NO3-) is reduced from 0.64 µg sm-3 to 0.12 µg sm-3 in the lowest 1 km. An increase in NH3 emissions in the LA region is critical to capturing ammonium nitrate (NH4NO3) formation down-wind of the city core. Using this modified emissions simulation, seasonal PM differences in the two regions and the export of excess NH3 out of the Valley are explored. Mean June simulated PM concentration in the lowest 1 km is 3.48 µg sm-3 near LA (38% NO3- and 39% SO42-, by mass) and 1.98 µg sm-3 in the Valley (44% NO3- and 32% SO42-, by mass). These simulated PM concentrations are 2 times greater in the Valley in December than in June, when NH4NO3 formation is favored by colder temperatures. However, LA simulated PM concentration falls by 53% in December compared to June, likely due to lower winter NH3 emissions. Both the model and IASI satellite observations indicate that large amounts of excess NH3 are transported from the Valley to southeastern California in the summertime which may negatively affect ecosystems in this area.
  • Thumbnail Image
    ItemOpen Access
    Assessment of numerical weather prediction model re-forecasts of atmospheric rivers along the west coast of North America
    (Colorado State University. Libraries, 2018) Nardi, Kyle M., author; Barnes, Elizabeth A., advisor; Schumacher, Russ S., committee member; Ham, Jay M., committee member
    Atmospheric rivers (ARs) - narrow corridors of high atmospheric water vapor transport - occur globally and are associated with flooding and maintenance of the regional water supply. Therefore, it is important to improve forecasts of AR occurrence and characteristics. Although prior work has examined the skill of numerical weather prediction (NWP) models in forecasting ARs, these studies only cover several years of re-forecasts from a handful of models. Here, we expand this previous work and assess the performance of 10-30 years of wintertime (November-February) AR landfall re-forecasts from nine operational weather models, obtained from the International Subseasonal to Seasonal (S2S) Project Database. Model errors along the West Coast of North America at leads of 1-14 days are examined in terms of AR occurrence, intensity, and landfall location. We demonstrate that re-forecast performance varies across models, lead times, and geographical regions. Occurrence-based skill approaches that of climatology at 14 days, while models are, on average, more skillful at shorter leads in California, Oregon, and Washington compared to British Columbia and Alaska. We also find that the average magnitude of landfall Integrated Water Vapor Transport (IVT) error stays fairly constant across lead times, although over-prediction of IVT is more common at later lead times. We then show that northward landfall location errors are favored in California, Oregon, and Washington, although southward errors occur more often than expected from climatology. We next explore the link between the predictability of ARs at 1-14 days and synoptic-scale weather conditions by examining re-forecasts of 500-hPa geopotential height anomaly patterns conducive to landfalling ARs. Finally, the potential for skillful forecasts of IVT and precipitation at subseasonal to seasonal (S2S) leads is explored using an empirical forecast model based on the Madden-Julian oscillation (MJO) and the quasi-biennial oscillation (QBO). Overall, these results highlight the need for model improvements at 1-14 days, while helping to identify factors that cause model errors as well as sources of additional predictability.
  • Thumbnail Image
    ItemOpen Access
    Climatology and variability of atmospheric rivers over the north Pacific
    (Colorado State University. Libraries, 2017) Mundhenk, Bryan D., author; Barnes, Elizabeth A., advisor; Maloney, Eric D., advisor; Randall, David A., committee member; Ham, Jay M., committee member
    Atmospheric rivers (ARs) are plumes of intense water vapor transport that dominate the flux of water vapor into and within the extratropics. Upon landfall, ARs are a major source of precipitation and often trigger weather and/or hydrologic extremes. Over time, landfalling AR activity, or a lack thereof, can influence periods of regional water abundance or drought. An objective detection algorithm is developed to identify and characterize these features using gridded fields of anomalous vertically integrated water vapor transport. Output from this algorithm enables the investigation into the relationships between tropical variability and ARs over the North Pacific undertaken in this dissertation. In the first segment of this study, an all-season analysis of AR incidence within the North Pacific basin is performed for the period spanning 1979 to 2014. The variability of AR activity due to the seasonal cycle, the El Nino-Southern Oscillation (ENSO) cycle, and the Madden-Julian oscillation (MJO) is presented. The results highlight that ARs exist throughout the year over the North Pacific. In general, the seasonal cycle manifests itself as northward and westward displacement of AR activity during boreal summer, rather than a seasonal change in the total number of ARs within the domain. It is also shown that changes to the North Pacific mean-state due to the ENSO cycle and the MJO may enhance or completely offset the seasonal cycle of AR activity, but that such influences vary greatly based on location within the basin. The second segment of this study investigates ARs at high northern latitudes. Comparatively little is known about the dynamics supporting these ARs in contrast to their mid-latitude counterparts. ARs are found to occur near the Gulf of Alaska and the U.S. West Coast with similar frequency, but with different seasonality. Composited atmospheric conditions reveal that a broad height anomaly over the northeast Pacific is influential to AR activity near both of these regions. When a positive height anomaly exists over the northeast Pacific, AR activity is often deflected poleward toward Alaska, while the U.S. West Coast experiences a decrease in AR activity, and vice versa. This tradeoff in AR activity between these two regions applies across a range of time scales, not just with respect to individual transient waves. Both ARs and height anomalies are found to be associated with Rossby wave breaking, thereby dynamically linking the modulation of AR activity with broader North Pacific dynamics. The third segment of this study explores the predictability of anomalous landfalling AR activity within the subseasonal time scale (approximately 2-5 weeks). An empirical prediction scheme based solely on the initial state of the MJO and the stratospheric quasi-biennial oscillation (QBO) is constructed and evaluated over 36 boreal winter seasons. This scheme is based on the premise that the MJO modulates landfalling AR activity along the west coast of North America within the subseasonal time scale by exciting large-scale circulation anomalies over the North Pacific. The QBO is found to further modulate the MJO--AR relationship. The prediction scheme reveals skillful subseasonal "forecasts of opportunity" when knowledge of the MJO and the QBO can be leveraged to predict periods of increased or decreased AR activity. Moreover, certain MJO and QBO phase combinations provide predictive skill competitive with, or even exceeding, a state-of-the-art numerical weather prediction model.
  • Thumbnail Image
    ItemOpen Access
    Decreasing nitrogen for volatilization in beef feedlot cattle
    (Colorado State University. Libraries, 2012) Kappen, Maria M., author; Archibeque, Shawn L., advisor; Engle, Terry E., committee member; Ham, Jay M., committee member
    The effects of ractopamine hydrochloride (RAC) and a steroidal implant (IMP), on whole body N metabolism were evaluated in 24 Hereford x Angus steers (BW 554.4 ± 26.8 kg). The experimental design was a completely randomized block design with a 2 x 2 factorial arrangement of treatments. Factors included: 1) RAC (0.0 or 400 mg×steer-1×d-1) and 2) IMP (0.0 or 200 mg trenbolone acetate and 28 mg of estradiol benzoate). Steers were housed in individual pens and allowed ad libitum access to feed and water throughout the experiment. Steers were acclimated to the metabolism barn by bringing in, tying and currying for 12 d before the initiation of the experiment. Once cattle had been implanted for 48 d and had received RAC for 21 d, a nutrient balance study was conducted for 6 d. An IMP x RAC interaction tended (P < 0.09) to exist for DMI. Implanted steers receiving RAC tended to have lower DMI compared to non-IMP steers receiving RAC as well as IMP steers not receiving RAC. N intake (P > 0.11) and fecal N (P > 0.18) were not different due to treatment, yet numerically reflected the trend noted for DMI. Urinary N excretion was decreased by feeding RAC (P < 0.01). There tended (P < 0.08) to be an IMP x RAC interaction for urinary N excretion. Implanted steers receiving RAC tended to have less urinary N than steers receiving an implant only. Similarly, urine urea N excretion was decreased by RAC treatment (P < 0.02) and excretion tended to be decreased in steers that had also received IMP (IMP x RAC interaction; P < 0.07). Overall N retention was not affected by treatment (P > 0.14). These results indicate that urinary N excretion can be reduced by incorporating RAC according to labeled usage during the final phase of the finishing period. However, more studies will be required to elucidate the potential interactions of RAC with implant status and types of implants.
  • Thumbnail Image
    ItemOpen Access
    Effect of organic nitrogen fertilizer source, application method, and application rate on ammonia volatilization from drip irrigated vegetables
    (Colorado State University. Libraries, 2018) Erwiha, Ghazala M., author; Davis, Jessica G., advisor; Ham, Jay M., committee member; Collett, Jeffrey L., committee member
    To view the abstract, please see the full text of the document.
  • Thumbnail Image
    ItemOpen Access
    Estimating emission rates of volatile organic compounds from oil and natural gas operations in the Piceance Basin
    (Colorado State University. Libraries, 2015) MacDonald, Landan Patrick, author; Pierce, Jeffrey R., advisor; Collett, Jeffrey L., advisor; Ham, Jay M., committee member
    Oil and natural gas production has been steadily increasing in Colorado for the past 10 years. Garfield County is partially located above the natural gas rich Piceance Basin. Horizontal drilling techniques provide increased access to subsurface gas deposits while hydraulic fracturing is employed to increase the permeability of the tight gas formations by pumping pressurized fluids into the ground to allow more cost-effective oil and gas extraction. Once fractured, the fluid is allowed to flow back to the surface to be captured before the well is considered producing. Our team conducted field measurements from 2013 to 2015 in Garfield County to determine emission rates of methane, hazardous air pollutants, and ozone precursors at 18 oil and gas operations. Drilling and well completion operations were targeted because they are understudied relative to production. We estimate the emission rates of methane and 58 additional VOCs (focusing on benzene, toluene, and ethane) for three common operations. We found benzene had mean emission rates of 0.72, 0.23, and 0.055 g/s for drilling, hydraulic fracturing, and flowback operations respectively. We calculated mean methane emission rates of 6.2, 29, and 64 g/s for drilling, hydraulic fracturing, and flowback operations respectively. We use the estimated methane emission rates from drilling and well completion operations to compare to typical well lifetime emissions during a 30 year production phase and find that drilling and well completion operations may be contributing from 0.1 to 10% of total well pad emissions. These results are based on a limited sampling size (18 sites) and limited overall measurement time (4.25 hours of total measurement time included in results). It is possible we did not perform measurements for long enough periods of time at enough sites. This study is beginning to fill the information gap by focusing on drilling and well completion operations. AERMOD is an atmospheric dispersion model used for new source apportionment. We compared our measured concentration fields to AERMOD predicted concentration fields by replicating fieldwork locations and conditions. Meteorological conditions were taken from an on-site meteorological station for use in the dispersion model. Comparing to the measurements, we found there was a low log-mean bias (-0.007) with a large amount of scatter (r = 0.0007). Additionally, we use AERMOD and data from the NCEP North American Regional Reanalysis database to predict the distribution of concentrations experienced throughout for various meteorological conditions in Garfield County at various distances surrounding oil and gas wells. We normalized these predicted concentration fields by emission rate and created cumulative distribution functions.
  • Thumbnail Image
    ItemOpen Access
    Evaluation of the Kipp and Zonen large aperture scintillometer for estimation of sensible heat flux over irrigated and non-irrigated fields in southeastern Colorado
    (Colorado State University. Libraries, 2012) Rambikur, Evan H., author; Chávez, José L., advisor; Andales, Allan A., committee member; Ham, Jay M., committee member; Gates, Timothy K., committee member
    The aim of this work was to assess the performance of the Kipp and Zonen Large Aperture Scintillometer (LAS; Delft, Netherlands) to predict surface sensible heat flux (H). The LAS was introduced approximately 30 years ago and has been marketed as an indirect tool for the estimation of vegetation evapotranspiration (ET). Several tests have shown the LAS to be a fairly robust tool for prediction of H, both over homogeneous and heterogeneous surfaces. However, the Kipp and Zonen LAS has been criticized for overestimation of H and for significant inter-sensor deviation in H. Field experiments were performed in 2011 using three Kipp and Zonen LAS units over two different surfaces to assess the accuracy and inter-sensor variability. Accuracy was evaluated based on reference measurements from eddy covariance (EC) instrumentation, which provides direct measurement of sensible and latent heat fluxes. Notably the EC method has been criticized for systematic underestimation of the sensible and/or latent heat flux, but is nonetheless a common tool used to validate LAS data. The first experimental test site was predominantly dry and uniform grassland located near Timpas, CO. At this site, all three LAS units were deployed together for some time in order to assess inter-sensor variability and an EC system was installed for some duration of the LAS deployment. The EC system was subsequently moved to the second site, which was the Colorado State University (CSU) Arkansas Valley Research Center (AVRC) near Rocky Ford, CO. At the AVRC, one LAS unit was set up over irrigated alfalfa. Results from the inter-LAS comparison suggested that there may be some inherent variability between 6-11% in LAS-predicted H (HLAS) and that the physical alignment of the LAS is critical for maintaining good performance. Testing different methods for estimation of the friction velocity (u*) variable revealed bias between the logarithmic wind profile (LWP) result and the EC measurement. Linear regression slopes between 0.94 and 1.35 were found for HLAS with respect to EC-derived H (HEC) for the Timpas site - dependent on the LAS unit, the LAS alignment, and the u* method. The overall conclusion was that HLAS was reasonably accurate, partially due to the potential of HEC being underestimated on the basis of lack of energy balance closure. For the CSU AVRC (irrigated) site, HLAS was generally observed to be greater than HLAS by 20-30%. However, heat flux source area differences between the LAS and EC units may have contributed to some of the observed biases. Further, the overall conclusion of reasonable accuracy of HLAS was made, again partially due to potential for H underestimation by the EC system. It is recommended, nonetheless, for future applications to calibrate the Kipp and Zonen LAS to a reliable reference on the basis of observed inter-sensor variability. Further, the benefit of the LAS is judged to be higher for a scenario of limited or no irrigation than for one of full irrigation, since the contribution of H to the overall energy balance would be relatively small for a full irrigation scenario.
  • Thumbnail Image
    ItemOpen Access
    Forest respiration from eddy covariance and chamber measurements under high turbulence and a bark beetle epidemic
    (Colorado State University. Libraries, 2013) Speckman, Heather, author; Ryan, Michael G., advisor; Parton, William J., advisor; Ham, Jay M., committee member; Denning, A. Scott, committee member
    Eddy covariance (EC) enables continuous estimates of carbon, water, and energy fluxes, and a global network of >500 sites (www.fluxnet.ornl.gov) has resulted in major advances in understanding ecosystem-scale biogeochemical cycling. However, long-term sums of net ecosystem exchange, photosynthesis and respiration fluxes have uncertainties because of potential measurement biases in respiration fluxes at night. Many studies have demonstrated that EC estimations of flux during the night are lower than chamber measurements--with low turbulence at night potentially causing the difference. A bark beetle outbreak at the GLEES Ameriflux site provided a unique opportunity to compare chamber and EC estimates of ecosystem respiration (R) under conditions of high turbulence (summer night mean u* = 0.7 m s-1) and 85% mortality of the aboveground respiring biomass due to a bark beetle epidemic. Chamber-based estimates of R were developed from periodic foliage, wood and soil CO2 efflux measurements fit to models of phenological seasonal change and diurnal temperature response. These models estimated ecosystem mean nightly respiration to have declined 32% after the bark beetle epidemic (7.0 ±0.22 μmol m-2 s-1 in 2005 to 4.8 ±0.16 μmol m-2 s-1 in 2011). The decrease was entirely due to the loss of aboveground respiration, soil efflux remained constant throughout the epidemic. Unlike chamber estimates, nighttime EC measurements did not decline after 85% of the forest basal area had been infested or killed by bark beetles, mean nighttime NEE of 3.0 μmol m-2 s-1 for 2005 and 2011. These EC values were significantly lower than chamber estimates of respiration for the same time periods (58% lower in 2005, and 34% in 2011). Despite the large difference in values, the two estimates of R were correlated (yearly r2 ranging from 0.18-0.60). This study suggests that the traditional discrepancy of nighttime EC and chamber estimates of ecosystem respiration are not caused by insufficient turbulence (results proved robust to extreme u* filter > 0.7 m s-1). Other sources of error are investigated for both techniques. To further explore this discrepancy, we suggest the installation of a second EC system below the canopy to improve understanding of air flows and fluxes throughout the ecosystem. This discrepancy must be resolved before scientific confidence can be attained in the true value of ecosystem carbon flux.
  • Thumbnail Image
    ItemUnknown
    Insights into the biosphere-atmosphere exchange of organic gases from seasonal observations over a ponderosa pine forest
    (Colorado State University. Libraries, 2020) Fulgham, S. Ryan, author; Farmer, Delphine, advisor; Ham, Jay M., committee member; Ravishankara, Akkihebbal R., committee member; Van Orden, Alan, committee member
    The biosphere-atmosphere exchange of organic gases over forests contributes to the formation of air pollution and the availability of forest nutrients. Forests can be both sources and sinks of volatile and semi-volatile organic compounds to the atmosphere. The role that forests play in controlling organic acid concentrations remains poorly understood, with multiple model-measurement comparisons reporting missing sources of formic acid. Large, missing sources of organic acids have been identified over different forested environments. Despite substantial seasonal variability in forest productivity and environmental conditions, a paucity of observations, during seasons other than summertime, is available. Although forest fires are a major source of hazardous organic gases and particulate matter, few measurements of semi-volatile organic compounds emitted by forest fires are available from within 1 km of the fire. Detection further-afield cannot disambiguate between chemistry at the source of the fire and chemical aging as a smoke plume traverses the atmosphere. Near-field observations are needed to characterize emissions attributable to combustion and pyrolysis processes. To improve understanding of processes that control the atmospheric budgets of organic acids, water-soluble pollutants with physicochemical properties similar to organic acids, and fire-emitted phenolic compounds, this dissertation reports measurements of the biosphere-atmosphere exchange of a suite of organic gases over a Rocky Mountain ponderosa pine forest in Colorado over four, seasonally-representative measurement campaigns. First, we report seasonally persistent, upward fluxes of organic acids, which are neither explained by direct emissions nor secondary production. Second, we present evidence for equilibrium partitioning into and out of water films on forest surfaces as both a missing source and sink of isocyanic acid and small alkanoic acids. Finally, we report significant enhancement of organic acids, phenolic compounds, and other nitrogen containing compounds during initiation of a controlled forest fire compared with the remainder of the burn. Nitrated phenols are rapidly produced and enhanced more than phenolic precursors during initial, higher temperature conditions. We attribute greater enhancement of nitrated phenols to high NOx emissions under higher temperature conditions.
  • Thumbnail Image
    ItemUnknown
    Management strategies for improved production practices to maximize efficiency associated with livestock production
    (Colorado State University. Libraries, 2014) Warner, Crystal M., author; Archibeque, Shawn L., advisor; Han, Hyungchul, committee member; Ham, Jay M., committee member
    With the ever-increasing world population of over 7 billion and subsequent increase in urbanization, it is crucial for the livestock sector of agriculture to move in the direction of sustainability. Appropriate changes in production practices ensure adequate production with fewer resources to meet the needs of the consumer. Multiple improvements within various management categories are essential to increase animal efficiency and economic gain, improved utilization of natural resources and reduce resulting environmental impacts. The National Air Quality Site Assessment Tool (NAQSAT), originally launched in 2010, provides its users the ability to qualitatively assess how effectively producers are mitigating harmful air emissions in site-specific beef, dairy, swine, broiler chicken, laying hen and turkey production facilities. The air emissions deemed to be of the greatest concern were odor, particulate matter (PM), ammonia (NH3), hydrogen sulfide (H2S), methane (CH4) and Volatile Organic Compounds (VOCs). Eight management categories are considered: animal housing, diet, manure handling, storage and application practices, mortality and road management. The tool enables users to run hypothetical scenarios to identify potential unintended consequences of management modification prior to making costly changes. The NAQSAT has since then been reviewed and updated by experts according to the most up-to-date knowledge and research to create version 2.0. The tool has expanded to include the horse species and air emission nitrous oxide (N2O). Following the implementation of the tool, users are directed to potentially applicable NRCS practices pertinent to their management goals within a given management category/emission of concern. Ideally this will guide users to reformed practices for continued sustainability in today's production environment. In order to accommodate this movement towards sustainability, diet modifications to the typical feedlot diet have been explored; 126 corn fed cross-bred steer calves (initial BW 529.5kg ± 10.7) were supplemented a rumen bypass fat during the last 60 days of the finishing period to evaluate its effects on feedlot performance, carcass characteristics and intramuscular fatty acid composition. Steers were blocked by initial, BW 9 head/pen (n = 7 pens / treatment), at the South Eastern Colorado Research Center (SECRC) in Lamar, CO. Pens were randomly assigned one of two treatment groups: 1) a control diet consisting of a regular corn based finishing ration (CON) and 2) rumen bypass fat treatment consisting of the control diet + Megalac-R/head/day (BF). Diets were formulated to be isocaloric and isonitrogenous. Animals were fed twice daily at 110% of the previous daily ad libitum intake. Feed bunks were cleaned and orts were collected weekly. Dry matter content was analyzed and diet samples were collected weekly for proximate analysis. Individual live weights were recorded and blood samples were collected on d -54, -10, 27 and 60 and 61. Feedlot performance and carcass characteristics were assessed (table 2). Initial BW was included in statistical analysis as a covariate. Steers fed the CON diet had a greater level of performance for most of the parameters measured; the CON treatment had greater DMI (10.14kg vs. 8.77kg; P<0.02) and tended to have greater ADG (1.699kg vs. 1.469kg;
  • Thumbnail Image
    ItemUnknown
    Micrometeorological studies of a beef feedlot, dairy, and grassland: measurements of ammonia, methane, and energy balance closure
    (Colorado State University. Libraries, 2018) Shonkwiler, Kira Brianne, author; Collett, Jeffrey L., advisor; Ham, Jay M., committee member; Kreidenweis, Sonia, committee member; Schumacher, Russ, committee member; Archibeque, Shawn, committee member
    Ammonia emissions from concentrated animal feeding operations (CAFOs; most of which are beef feedlots) near the Colorado Front Range are suspected to be a large regional input of reactive nitrogen which has been found to accumulate and cause deleterious effects in nearby downwind Class I areas like Rocky Mountain National Park. Methane (CH4) is a strong greenhouse gas (GHG) emitted in large amounts from dairy anaerobic lagoons used for liquid manure management. Lagoon systems account for over half of the manure management-based CH4 emissions from agriculture in the US. There is a strong need for more emissions measurements from CAFOs like feedlots and dairies. For these data to be trusted, well-developed techniques must be utilized at emissions measurement sites and such techniques should be validated in ideal scenarios. Three micrometeorological studies were performed involving measurement of emissions using micrometeorological methods in the surface layer. The first study involved estimating summertime NH3 emissions from a 25,000-head beef feedlot in Northern Colorado. Two different NH3 sensors were used: a cavity ring down spectroscopy analyzer collected data at a single point while a long-path FTIR collected data along a 226-m long transect, both deployed along the same fenceline. Concentration data from these systems were used with two inverse dispersion models (FIDES, an inverse solution to the advection dispersion equation; and WindTrax, a backward Lagrangian stochastic model). Point sensor concentrations of NH3 were similar to line-integrated sensor concentrations suggesting some spatial uniformity in emissions. Emissions had a diurnal pattern (i.e., afternoon peak with minimum in early morning) that was driven by temperature. Emissions predicted by WindTrax were 25.2% higher than those from FIDES. Point vs. long-path measurements of NH3 had minimal effect on predicted emissions. The mean NH3 emission factor (EF) was 80 ± 39 g NH3 hd−1 d−1, with 40.0% of dietary-N emitted as NH3. The second study involved using eddy covariance and WindTrax to quantify CH4 emissions from a 3.9-ha anaerobic lagoon serving a 1400-head dairy in northern Colorado. Methane emissions followed a strong seasonal pattern correlated with temperature of the organic sludge layer on the bottom of the lagoon. Fluxes started increasing in late spring (May; ~10°C), increased rapidly in Jun (10-15°C) peaked in the summer (Jul/Aug; ~18-20°C) and remained high until mid-autumn (late Oct/early Nov; ~10°C). Fluxes then decreased and remained consistently low (up to 10 times less than peak emissions) until microbial activity ramped up again in May. The EC signal was very dependent on wind direction, with highest concentrations and fluxes associated with the direction of the lagoon. Gap-filled data showed a slight diurnal pattern to all seasons, with tenfold increases in diurnal values for summer over winter. Additionally, EFs for the lagoon varied by season with lows in the winter and highs in the summer with an annual mean of 819 ± 774 g CH4 hd-1 d-1. WindTrax overestimated EC for the lagoon (1163 ± 1049 g CH4 hd-1 d-1 versus 819 ± 774 g CH4 hd-1 d-1), but this difference may be attributable to differences in the sampling footprint and stability conditions. IPCC Tier 2-calculated EFs were extremely close to EC-based measurements and WT-based estimates. The third study involved using eddy covariance in an ideal environment (tallgrass prairie in Kansas) to test the reasons behind the "energy balance (EB) closure problem" at two landscape positions. This problem can cast uncertainty on flux measurements made by EC. One upland and one lowland EC tower each were used to measure EB components (i.e., net radiation, Rn; soil heat flux, G; total change in heat storage, deltaS; and sensible and latent heat fluxes, H and λE) during the summers of 2007 and 2008. To maximize closure, special attention was given to reduce all forms of instrumentation error and account for heat storage and photosynthesis between the soil and the reference height. Landscape position had little effect on G, H, and Rn; differences were ≤ 2% between sites. Lowland λE was 8% higher than upland λE because of greater biomass and soil moisture. On average, EB closure (i.e., Σ[λE+H] / Σ[Rn–G–ΔS]) was 0.88 and 0.94 at the upland and lowland sites, respectively. Closure was not correlated with friction velocity or the stability of the surface boundary layer. Given high confidence in Rn, G, and ΔS, turbulent fluxes depend directly on vertical velocity (w), and the fact that a systematic underestimation of w was recently found in literature, lack of closure may have resulted largely from anemometer-based underestimates of w.
  • Thumbnail Image
    ItemUnknown
    Modeling sensible heat flux for vegetated surfaces through an optimized surface aerodynamic temperature approach
    (Colorado State University. Libraries, 2019) Costa Filho, Edson, author; Chavez, Jose L., advisor; Ham, Jay M., committee member; Venayagamoorthy, Karan, committee member
    Agricultural water management advancements rely on improved methods to accurately determine crop water use. Crop evapotranspiration modeling based on the surface energy balance depends on the accurate estimation of all incoming and outgoing heat fluxes at the surface level. This thesis particularly goal is to improve sensible heat flux estimates for row crops through an optimized aerodynamic surface temperature (To) approach based on remote sensing and weather data. Empirical linear and non-linear To models were developed based on percent cover, surface temperature, air temperature, and a new variable named turbulent mixing row resistance using data collected at the USDA-ARS Research Farm located in Greeley (CO). The experiment took place in two sub-surface drip irrigation corn fields with different irrigation water management practices in 2017-2018. Sensible heat flux were measured using LAS, eddy covariance, aerodynamic profile, and Bowen ratio methods. Remote sensing data were measured on-site using a radiometer. The fields were considered a point in space. Data from Aimes (IA) and Rocky Ford (CO) were used to assess proposed model performances under different locations and in comparison to published To models. The results have indicated that the optimized linear To models performed better than the non-linear and published models approaches, indicating that the introduction of percent cover and the new variable has provided reliable results under different data sets. The linear proposed To approaches improved sensible heat flux estimation, on average, in 33 % and 28 % for the deficit and fully irrigated field at LIRF in comparison to the sensible heat based on published To models. Sensible heat flux modeling results were better for the modeling approaches considering the empirical linear To model than the non-linear approaches for all three data set tested.
  • Thumbnail Image
    ItemUnknown
    Predicting water content and saturation in mine tailings with an electromagnetic soil moisture sensor
    (Colorado State University. Libraries, 2023) Martin, Garret M., author; Bareither, Christopher A., advisor; Scalia, Joseph, IV, committee member; Ham, Jay M., committee member
    The degree of saturation of mine tailings plays an important role in geotechnical and geochemical stability of a tailings facility, and as such, reliable measurements of in situ tailings saturation aid in evaluating the stability of a tailings facility. However, measuring in situ saturation in tailings facilities is a common challenge in the tailings industry. The objectives of this study were to (1) evaluate the ability of an electromagnetic soil moisture sensor to predict the volumetric water content and degree of saturation of mine tailings and (2) conduct proof-of-concept tests to assess the potential for electromagnetic sensors to be used as a tool in tailings engineering practice. To meet these objectives, laboratory-scale testing was conducted using an electromagnetic soil moisture sensor embedded in moist-tamped and slurry-deposited specimens of a single hardrock mine tailings prepared at varying volumetric water content, degree of saturation, and dry density. Certain specimens were subjected to changes in mass-volume properties and sensor performance was evaluated for timeliness and accuracy of response. The results of this study indicate an electromagnetic soil moisture sensor can be used to predict the volumetric water content and degree of saturation in hardrock mine tailings with a useful degree of accuracy depending on the application and precision required. During the proof of concept tests performed, error in predicted volumetric water content was less than about 1.5% to 3.8%, error in predicted saturation was predominantly less than 5%, and temporal response to changes in moisture was equivalent to a sensor insertion rate of approximately 27 mm/s. Based on the findings of this study, electromagnetic sensor technology offers a viable tool to predict the degree of saturation within tailings facilities and can be incorporated into innovative approaches to address the challenges encountered in different types of tailings facilities.
  • Thumbnail Image
    ItemUnknown
    The effect of crude protein withdrawal and the use of oscillated crude protein concentration on feedlot performance, carcass merit, and ammonia emissions from the pen surface of feedlot steers
    (Colorado State University. Libraries, 2011) Westover, Elin C., author; Wagner, John J., advisor; Engle, Terry E., advisor; Ham, Jay M., committee member
    Six hundred crossbred steers (BW, 329.7 ± 7.58 kg) were used to investigate the effect of CP withdrawal and the use of oscillating CP concentrations on feedlot performance, digestibility, carcass merit, and ammonia emissions from the pen surface of yearling steers. Steers were randomly assigned to one of the following treatments: HCP [Control, 13.5% CP, 3.5% CP equivalents (CPE)]); OCP (11.62% CP, 1.5% CPE fed Wednesday, Thursday, and Sunday and the HCP diet fed Monday, Tuesday, Friday, and Saturday); EICP (12.56% CP from d28 to slaughter, 2.53% CPE); ELCP (11.62% CP from d28 to slaughter, 1.55% CPE); LICP (HCP throughout with the ICP diet fed the last 27d); and LLCP (HCP throughout with the LCP diet fed the last 27d). Urea was used to modify dietary CP concentrations. Steers were housed in 9-steer pens (n=48) or 7-steer mass balance pens (n=24). Steers were weighed and ultrasound images and fecal grab samples were taken 3 or 4 times (depending upon replicate) throughout the trial. Feed samples, fecal grab samples, and mass balance pen surface samples were analyzed for DM, AIA, N, and P. Soil samples were obtained from the mass balance pens for treatments HCP, OCP, and ELCP and tested for total ammonia volatilization. Steers were harvested on d 149 or d 175 and camera carcass data was collected. Although initial BW differences between treatments were not significant (P > 0.18), initial BW was a significant (P < 0.10) source of variation describing interim and final BW and was therefore included in the data analysis as a covariate. There were no treatment differences for BW (P > 0.23) throughout the study. Average daily gain for each time period or for the entire study was not affected by treatment (P > 0.26). There was a difference (P < 0.05) in DMI between treatments from d 106 to slaughter (HCP > ELCP, LLCP, and LICP), and overall DMI tended (P < 0.11) to be affected by treatment (HCP > ELCP and LLCP). Treatment differences for G: F and net energy recovery were not significant (P > 0.30). There were no significant (P > 0.21) effects of dietary treatment on carcass merit. Treatment differences for DM digestibility calculated from DMI and fecal output as estimated by AIA, were not significant (P > 0.37) and averaged 85.7, 83.6, 84.2, and 83.0% for the HCP, OCP, EICP, and ELCP diets respectively. Treatment differences for CP digestibility, calculated from N intake and fecal N, were significant (P < 0.001) and averaged 83.3, 76.6, 78.8, and 74.3% for the HCP, OCP, EICP, and ELCP diets respectively (HCP > OCP, EICP, and ELCP). Nitrogen intake was significantly (P < 0.0001) affected by treatment and averaged 183, 172, 167, and 155 g per head daily for the HCP, OCP, EICP, and ELCP treatments respectively. Differences between treatments for amount of fecal N (P > 0.18) and calculated amount of retained N (P > 0.42) were not significant. Urinary N, calculated as N intake minus fecal and retained N, excretion was reduced (P < 0.0001) as N intake decreased with treatment averaging 128, 111, 108, and 94 g per steer daily for the HCP, OCP, EICP, and ELCP treatments, respectively. Retained N as a percentage of N intake increased (P < 0.0001) and calculated urinary N excretion decreased (P < 0.001) with decreasing N intake associated with treatment averaged 12.6, 13.6, 14.1, and 15.0% and 69.8, 64.6, 64.6, and 60.7% of N intake for the HCP, OCP, EICP, and ELCP treatments respectively. Cattle on the ELCP diet had significantly lower N loss than the HCP treatment (P < 0.02) and the OCP treatment (P < 0.10) for sampling from d 45 and d 92. Similar results were observed from samples taken on d 148; however there were no significant differences. Ammonia flux reduction of ELCP diet compared to HCP diet decrease from 40% to 21% with increasing days on feed. The average ammonia flux over the feed period for all treatments was 147.3 g/m2/d. There were no treatment differences (P > 0.36) for N, P, or N: P ratio found in samples from manure cleaned from the pen surface at the end of the study. Nitrogen to P ratios ranged from 2.13 to 2.23 and was lower than the fecal grab sample N: P ratio. These results indicate that ADG and carcass merit were similar for steers fed OCP and CP withdrawal diets as compared with the HCP control. Although DMI declined during the later stages of the finishing period, feed efficiency was not impacted by OCP or CP withdrawal diets. Reduced CP intake whether it was through the OCP or CP withdrawal diets was associated with less urinary N excretion and lower ammonia emissions from the pen surface.
  • Thumbnail Image
    ItemOpen Access
    The relative influence of aerosols and the environment on organized tropical and midlatitude deep convection
    (Colorado State University. Libraries, 2014) Grant, Leah Danielle, author; van den Heever, Susan C., advisor; Johnson, Richard H., committee member; Ham, Jay M., committee member
    In this two-part study, the relative impacts of aerosols and the environment on organized deep convection, including tropical sea-breeze convection and midlatitude supercellular and multicellular deep convection, are investigated within idealized cloud-resolving modeling simulations using the Regional Atmospheric Modeling System (RAMS). Part one explores aerosol-cloud-land surface interactions within tropical deep convection organized along a sea breeze front. The idealized RAMS domain setup is representative of the coastal Cameroon rainforest in equatorial Africa. In order to assess the potential sensitivity of sea-breeze convection to increasing anthropogenic activity and deforestation occurring in such regions, 27 total simulations are performed in which combinations of enhanced aerosol concentrations, reduced surface roughness length, and reduced soil moisture are included. Both enhanced aerosols and reduced soil moisture are found to individually reduce the precipitation due to reductions in downwelling shortwave radiation and surface latent heat fluxes, respectively, while perturbations to the roughness length do not have a large impact on the precipitation. The largest soil moisture perturbations dominate the precipitation changes due to reduced low-level moisture available to the convection, but if the soil moisture perturbation is moderate, synergistic interactions between soil moisture and aerosols enhance the sea breeze precipitation. This is found to result from evening convection that forms ahead of the sea breeze only when both effects are present. Interactions between the resulting gust fronts and the sea breeze front locally enhance convergence and therefore the rainfall. Part two of this study investigates the relative roles of midlevel dryness and aerosols on supercellular and multicellular convective morphology. A common storm-splitting situation is simulated wherein the right-moving storm becomes a dominant supercell and the left-moving storm evolves into a multicellular cluster. The right-mover, which is a classic (CL) supercell in the control simulation, becomes a low-precipitation (LP) supercell with increasing dryness aloft. Different midlevel hail growth mechanisms are found to dominate in the CL and LPs that assist in explaining their varying surface precipitation distributions. Although the CL and LP supercells are dynamically similar, their microphysical structure differs due to the strong control that midlevel dryness exerts on supercell morphology; aerosols have little impact on the supercellular structure. On the other hand, while midlevel dryness also dominates the changes to the multicellular convection, aerosols influence the precipitation through feedbacks to the cold pool strength and subsequent dynamical forcing. Overall, aerosol impacts are largest for the most weakly organized convection (tropical sea breeze convection) and smallest for strongly dynamic convection (supercells). Additionally, aerosol impacts are modulated by environmental influences, most notably soil moisture availability and midlevel moisture content in this study.
  • Thumbnail Image
    ItemOpen Access
    Unsaturated fluid flow and volume change behavior of filtered tailings
    (Colorado State University. Libraries, 2022) Aghazamani, Neelufar, author; Scalia, Joseph, IV, advisor; Bareither, Christopher A., committee member; Shackelford, Charles D., committee member; Ham, Jay M., committee member
    As the global demand for minerals continues to increase so does the generation of mine waste. Tremendous volumes of mine waste, viz. tailings and waste rock, are generated and placed in impoundments and piles. Improved methodologies are needed to enhance the sustainability of mine waste management. Tailings are typically discharged as a slurry of finely ground rock within water to a settling facility contained by an embankment. These facilities often necessitate long-term management of inherently weak materials. Tailings dewatered via filtration yields enhanced water stewardship and greater geotechnical stability; filtered tailings are readily amenable to progressive closure and environmental restoration. But, the high cost of tailings filtration, and the potential for acid rock drainage (ARD) due to oxygen ingress and internal unsaturated flow of water have limited the adoption of filtered tailings by the mining industry. The goal of this study is to advance the state of knowledge of filtered tailings. To this end, this dissertation consists of three components; (1) assessing the influence of filtered tailings placement conditions on filtered tailings unsaturated characteristics, (2) assessing excess pore pressure generation during compression of unsaturated filtered tailings, and (3) exploring hydrologic paths to minimize ARD and improve geochemical stability of filtered tailings stacks. Pressure plate, chilled mirror, and shrinkage curve tests were performed to produce soil water characteristic curves (SWCCs) for two precious metal mine tailings with varying initial water contents and dry densities. The resultant SWCCs illustrate that the placement water content and dry density have a significant influence on the unsaturated characteristics of the filtered tailings. Generation of excess pore water pressure was assessed via undrained compression tests. Unsaturated filtered tailings started generating appreciable excess porewater pressure (> 10% of the incrementally applied total vertical stress) when the saturation of the tailings was at the range of 80 to 90%; this appreciable excess pore pressure did not fully dissipate after 24 h. Filtered tailings in this study followed a stress path with void ratios below the critical state line (i.e., dilative during undrained shear) unless placed initially loose and wet. The SWCCs produced were used to model the hydrology of filtered tailings and comingled filtered tailings columns via HYDRUS-2D in example wet, hemiboreal, and dry climates. Results of this study illustrate that for the filtered tailings evaluated in this study, ARD is anticipated to be minimized via varying climate-dependent mechanisms. In the arid climate, filtered tailings functioned as a water balance layer, storing, and releasing precipitation as evaporation not percolation. In the wet climate, filtered tailings became rapidly saturated and maintained a saturated surficial layer preventing inward movement of oxygen and potentially minimizing generation of ARD. In the hemiboreal and wet climates percolation was minimal due to the low hydraulic conductivity of the filtered tailings, and inclusion of a commingled capillary barrier layer further reduced percolation (further reducing the potential flux of ARD). Results from this study illustrate the potential efficacy of filtered tailings to maintain both geotechnical stability and limit ARD.
  • Thumbnail Image
    ItemOpen Access
    Using modelling tools to advance the understanding of ammonia dry-deposition and bidirectional flux processes next to large animal feeding operations
    (Colorado State University. Libraries, 2020) Lassman, William, author; Pierce, Jeffrey R., advisor; Collett, Jeffrey L., Jr., advisor; Fischer, Emily V., committee member; Ham, Jay M., committee member
    Ammonia in the atmosphere is a trace gas that can play a big role in the Earth's climate, as well as human and ecological health. Due to its stickiness and solubility, ammonia can enter the biosphere via wet and dry deposition, where excess ammonia input often results in soil acidification, disruption of natural ecological equilibria, and loss of biodiversity. Additionally, ammonia is the most abundant alkaline species in the atmosphere and can react with atmospheric acids to form aerosols, which can affect the earth's radiative balance as well as human health. Ammonia emissions tend to be associated with agricultural sources, such as fertilized fields or animal waste at concentrated Animal Feeding Operations (CAFOs). Consequently, ammonia emissions tend to be dynamic and highly heterogeneous, and ammonia surface-fluxes are difficult to measure. However, in regions with many large CAFOs, ammonia can be an important regional pollutant, especially if there are sensitive ecosystems or other regional sources of atmospheric acids present. In this dissertation, I study ammonia dry-deposition fluxes immediately downwind of CAFOs using a variety of modelling tools. First, I discuss original research where I use a coupled a K-epsilon model with a Lagrangian-Stochastic ammonia bidirectional exchange surface model to simulate the dispersion and deposition of ammonia downwind of an idealized CAFO. Based on these simulations, the amount of ammonia that undergoes dry deposition depends greatly on the land surface downwind of the CAFO; replacing bare soil or unmanaged grassland with leafier surfaces such as cropland or forests can increase the fraction of total ammonia emissions that deposits from 2 - 10% to 30 - 50%, though this is sensitive to the ammonia emission potential in the model plant canopy. Next, I describe a separate study where I use a 3-D Large-Eddy Simulation model to simulate the dispersion of ammonia and methane from a CAFO with a time-resolved modelling tool. I use this modelling system to produce synthetic observations, which are used to develop an inversion approach to quantify the ammonia dry deposition near a CAFO with colocated mobile measurements of ammonia and methane. While I demonstrate that such an inversion technique is feasible with surface-based measurements, considerable value is added, in terms of minimizing method bias and increasing method precision, by mounting measurements on a small Unmanned Aerial System (sUAS). Finally, I present measurements of PM2.5 concentration and composition that were made in Palapye, Botswana. Botswana is a developing country with a hot and arid climate. Beef and livestock production are important economic activities in Botswana; however, the agricultural practices differ considerably from the CAFOs discussed in the rest of the dissertation. Furthermore, these livestock activities occur against a backdrop of emissions and air pollutants that differ considerably from the United States and Europe. The measurements show that PM2.5 concentrations were on average 9 μg m-3 during the 5-week measurement period. While below levels that are typically considered hazardous, there was considerable variability in the measured concentrations, and the measurement period is too short to conclusively determine that air pollution is not a public health concern in this region. The aerosol composition is dominated by carbonaceous species, probably from biomass burning, though inorganic sulfate also is abundant in the aerosol phase. As Botswana continues to undergo economic development, the types of emissions and pollution present will continue to change.
  • Thumbnail Image
    ItemOpen Access
    Volatile organic compound and methane emissions from well development operations in the Piceance Basin
    (Colorado State University. Libraries, 2016) Hilliard, Noel G., author; Collett, Jeffrey L., advisor; Fischer, Emily V., committee member; Ham, Jay M., committee member; Hecobian, Arsineh, committee member
    The natural gas industry in Colorado has experienced significant growth in the last decade due to widespread use of unconventional natural gas extraction technologies. Garfield County is located in the Rocky Mountain Region on the western slope of Colorado above the Piceance Basin. Natural gas wells in this region penetrate the William’s Fork formation, located approximately 4,000 ft. below the surface, which is a tight sand formation known to be rich in natural gas. Horizontal drilling increases the extraction potential of natural gas stored in several sandstone lenses. Hydraulic fracturing is a stimulation technique used to maximize the flow and efficiency of natural gas transport to the surface from unconventional reservoirs. Once the formation is adequately cracked, 10-50% of the hydraulic fluid flows back to the surface . Our field team collected samples in Garfield County between 2013-2015 to measure methane, ozone precursors, and air toxics associated with natural gas extraction activities. Very few studies have provided direct observations of VOC emissions from individual well development activities. Emission rates of 48 VOCs and methane were determined using the tracer ratio method for three well development operations: drilling, hydraulic fracturing (fracking), and flowback for a subset of samples collected. Methane had mean emission rates of 1.57, 6.78, and 25.6 g s-1 for drilling, hydraulic fracturing, and flowback operations respectively, while toluene had mean emission rates of 1.24, 0.469, and 0.437 g s-1 for these operations. Measured emission rates were used to determine if specific VOCs were well correlated with each other and/or methane emission rates. Strong correlations between individual VOC emission rates and methane were investigated to assess whether methane emission rates might serve as useful surrogates for emission rates of individual VOCs, which are less easily measured. We found that methane and ethane appear to be emitted from the same sources for all operation types indicating that methane emission rates may be useful surrogates for ethane emission rates. Methane emission rates appear not to be very useful surrogates for heavier VOCs, including C5-C10 alkanes, alkenes, and aromatics. Concentration ratios of source-specific tracer compounds were investigated to determine the source signatures of individual operation types. We found that drilling emissions appear to be primarily influenced by combustion, while flowback emissions are primarily influenced by the release of natural gas and other substances from the well.