Airborne data from the 2017 SARP DC-8 flights over the Las Angeles Basin, the Inland Empire, and Bakersfield indicate abnormally high concentrations of α-pinene across the region. The concentrated presence of α-pinene directly affects air quality by means of photochemically producing tropospheric ozone and secondary organic aerosols. While α-pinene has many sources, certain characteristics point towards a biogenic origin. During these flights, a regional heat wave with temperatures of over 105° F in some areas may be to blame. A relatively strong correlation between the levels of isoprene, β-pinene, and α-pinene may be indicators that these elevated concentrations resulted from heat stress. All three of these molecules are classified as Biogenic Volatile Organic Compounds (BVOCs) and have been shown in laboratory conditions to be elevated at higher temperatures. It is also important to note that every species emits different compounds at different levels under different conditions. This implies that the emissions from some plants (such as Eucalyptus and Sweetgum trees) may have a more direct impact on air quality than others. By running back trajectories of air masses with high concentrations of α-pinene, some areas and their relationship to the climate are brought into question. However, only a handful of data sets encompassing large areas over the course of years currently exist. Our understanding of large-scale or regional patterns of emission by biological sources is limited. This makes long-term SARP data a valuable resource for clarifying assumptions about plant emissions. More research into this topic could help to better our understanding of interactions between natural and urban landscapes. With an anticipated global increase of temperature in the future, these events which increase BVOC emissions may become more frequent. To minimize environmental impacts and health risks, more resilient plants with lower emission rates may be better suited for urban environments.

Studies find national methane emissions from the EPA are underestimated by 50%, and fossil fuel emissions are approximately 5 times smaller than other estimates (Miller et al, 2013). Thus, as the U.S. attempts to quantify or reduce greenhouse gas (GHG) emissions, they grossly underestimate current levels. Airborne measurements can cover a large area with precise measurements of GHGs. However, because SARP DC-8 flights measure 1000-2000 feet above ground level, it is not apparent where elevated concentrations originate. This research combines a dispersion model called STILT (Stochastic Time Inverted Lagrangian Transport), chemical ratios of tracer gases, and airborne measurements to find the potential locations and types of oil spills. Spills contaminate the water and soil, and leak gases that contribute to ozone formation and poor air quality (Evens & Helmig, 2017). In 2012, ethane levels in a rural, southern area in the San Joaquin Valley were above 8000 ppt, more than double the annual high from other years. Other years also have elevated ethane measured in the same area. Levels are comparable to Bakersfield, which has high, expected urban ethane sources. Enhanced levels of ethane may indicate a fossil fuel leak and often elevated methane emissions. Ratios of isopentane to n-pentane indicate whether a source is likely crude or refined oil (Gilman et al, 2013). This ratio is low in the southwestern San Joaquin Valley from 2012 onward, indicating a strong crude oil emissions signal. The STILT model shows that air from these samples is not enhanced by any Bakersfield emissions; it has a separate source rivaling a large urban area. The model points to a few oil fields as potential sources of leaks, demonstrating the potential to combine chemistry, modeling, and airborne measurements to find unreported oil leaks, spills, or dumping. This helps monitor regional impacts and national methane emissions.

Dimethyl sulfide (DMS) is a compound that is primarily produced from marine-based sources and is extremely common in and over the ocean. Analysis of the previous ten years of whole air samples collected during NASA’s Student Airborne Research Program (SARP) flights revealed elevated levels of DMS in the northern and southern Central Valley during the SARP flights in 2009, 2010, and 2014. Although the samples were collected in the middle of the Central Valley, DMS mixing ratios were comparable to expected levels in air over the ocean. Two different backwards trajectory models predicted that the DMS from those specific SARP years likely came from the Pacific Ocean, specifically areas near the San Francisco Bay or farther north. The presence of marine tracers (eg. CH3I, CH3ONO2) in the Central Valley confirms that the Pacific Ocean is the likely source of the DMS in the Central Valley. Other sources, such as fires, dairies, and inland bodies of water were also considered. Their contribution, if any, is very small based on back trajectories, wind direction, trace gases, and the timing of wildfires in relation to the SARP flights. The back trajectories indicated areas of the ocean with kelp forests and plankton blooms as possible sources, both of which produce large quantities of DMS. These high levels of DMS seen in the Central Valley are concerning because DMS undergoes reactions in the atmosphere and ultimately produces sulfate, which can form aerosols. These aerosols contribute to the Central Valley’s worrisome air quality and are a setback for the efforts to improve the air quality in the Central Valley since this air is being imported from the ocean.

Increasing interest in hydrogen (H2) as a fuel source has created the need to study the global atmospheric hydrogen budget and potential implications of hydrogen use. Hydrogen is a low carbon fuel source and is an attractive clean energy alternative to fossil fuels, as its combustion releases only water vapor and heat. Although hydrogen fuel cell vehicles have low emissions, there is agreement that there will be some leakage of hydrogen into the atmosphere during the storage and transportation process. It is unclear what the leakage rates will be, with estimates ranging from 2-10%. This leakage can have several negative consequences. An increased concentration of hydrogen in the atmosphere may lead to increased water vapor in the stratosphere, stratospheric cooling, and increased polar stratospheric clouds which can contribute to stratospheric ozone depletion. Whole air samples collected onboard the NASA Sherpa C-23 during the 2017 Student Airborne Research Program show elevated levels of H2 near LAX (920 ± 160 ppb) and Bakersfield (710 ± 80 ppb) when compared with the global average atmospheric mixing ratio of H2 (510 ppb). NOAA HYSPLIT back trajectories indicate elevated levels of hydrogen in the LAX area were likely in part from hydrogen refueling stations in the area. Slightly elevated levels of hydrogen were also observed in Bakersfield with no hydrogen refueling stations nearby. HYSPLIT back trajectories indicate surrounding airports may have been the source of the elevated hydrogen in Bakersfield. This suggests that airports may be contributing to the levels of hydrogen on a local and possibly global scale. This project highlights the importance of a better understanding of the global hydrogen budget and can give new insight into factors that contribute to sources of hydrogen in the atmosphere.

Chlorodifluoromethane, known as HCFC-22, is a hydrochlorofluorocarbon (HCFC) used as a propellant and refrigerant in industrial and residential applications. HCFCs were created to help phase out the use of chlorofluorocarbons, a class of gases extremely efficient at depleting stratospheric ozone. Even so, HCFC-22 is still an extremely potent greenhouse gas that has potential to deplete some stratospheric ozone. Mandated by the Montreal Protocol, the United States has begun to phase out HCFC-22. As of 2010, HCFC-22 can only be produced or imported to service extant air conditioning systems, banning usage in new systems. The NASA Student Airborne Research Program has collected whole air samples over Southern California since 2009, enabling analysis of trends in HCFC-22 relative to background levels. There is a significant difference in HCFC-22 levels between the Los Angeles Basin and the Inland Empire, with higher concentrations found in the latter. Local concentrations of HCFC-22 have not changed significantly in the Los Angeles Basin, whereas local concentrations in the Inland Empire increased until 2014 and have decreased since. Residential housing trends are an important factor in determining regional HCFC-22 emissions due its use in residential HVAC systems. The phaseout of HCFC-22 coincided with the national housing crisis, meaning that the earlier air samples reflect emissions from homes built before 2010, and the later air samples reflect additional emissions from homes built after 2010. The higher rates of new home construction after 2010 in the Inland Empire correspond to decreased use and emissions of HCFC-22 after 2014, where no significant change in home construction in Los Angeles Basin correspond to relatively static concentrations of HCFC-22 in the area. This shows that the Montreal Protocol regulations of HCFC-22 in the United States have been effective at reducing local residential emissions.

Earthquakes have the potential to create openings for trace gases to escape from the Earth’s subsurface. Seismic activity from earthquakes create pathways for gases to escape through fractures of the rocks within the faults. A recent earthquake near Ridgecrest, California with a magnitude 7.1 occurred July 5, 2019, which created migration pathways for trace gases such as SO₂ and CO₂ that may be emitted from the activated fault lines. In situ atmospheric trace gas measurements were taken with AutOMobile Trace Gas Surveyor (AMOG) on July 6, 2019 within twenty-four hours of the earthquake. Literature suggests that after an earthquake occurs, trace gases will remain elevated throughout the duration of aftershocks and will cease when the quaking stops. Ground samples in Searles Valley, Panamint Valley, and Death Valley were collected using whole air sampling and AMOG techniques. An analysis of airborne whole air samples aboard the DC-8 from SARP 2019 near Ridgecrest showed slightly elevated levels of CO₂ and SO₂. Just like the gases measured with AMOG, it is possible that these gases were emitted during or after the earthquakes. During an earthquake, local possible sources like Death Valley’s Ubehebe crater could be sources of SO₂. Based upon the collected data, one can speculate that the trace gases are coming out of rock fractures, however there is not enough evidence to confirm that the faults are the actual source of gas emissions. Further ground level and airborne studies will provide data that will enhance our knowledge between earthquakes and gaseous emissions.

Ground-level whole air samples were collected upwind, downwind, and next to E & J Gallo Winery in Fresno, California. E & J Gallo is the largest winery in Fresno, and the largest privately-owned wine company in the world. These samples were analyzed using gas chromatography and a variety of detectors. Elevated levels of ethanol were found downwind of the winery. These levels alone are not room for concern; however, the products of atmospheric ethanol oxidation may create problems. Ethanol oxidation creates acetaldehyde, which can increase the levels of ozone and secondary organic aerosols. Forward trajectories indicate that these emissions travel south through San Joaquin Valley through areas that California has labelled as “Disadvantaged Communities.” These are areas known for poor air quality, among other issues, that the state is trying to address. The level at which these gases are emitted could have negative health and quality of life effects for locals in this area, especially over a lifetime. Although the concentrations were not above the toxicity thresholds for these gases, there are hundreds of wineries scattered throughout these Disadvantaged Communities. These levels might be even worse later in the year. Harvested grapes will be destemmed and crushed for later use; emissions during this time are predicted to be much higher than what we found. This step is believed to have the largest ethanol emissions during the wine production process, which in turn creates large amounts of acetaldehyde and ozone. The crush is not until later in the year, but these ground samples can be compared to SARP 2019 and FIREX-AQ flights. Later, a comparison could be drawn between policymakers’ calculations and observed emissions. For now, this research examines what impact wineries may have on the Disadvantaged Communities in California.

From 2012 to 2016 California experienced an extreme drought during which conifer mortality increased significantly. The United States Forest Service (USFS) estimated that there were 149 million dead trees in California by the end of 2018. This study specifically focuses on the Figueroa Mountain area, which is located in the Los Padres National Forest. A study region was determined in the Los Padres National Forest based on the extent of the 2019 Airborne Visible / Infrared Imaging Spectrometer Next Generation (AVIRIS-NG) images. In order to identify conifer mortality in the study region, two AVIRIS-Classic flight lines for 2013 and 2015 (18 meter spatial resolution) and two AVIRIS-NG flight lines for 2019 (8.6 meter spatial resolution) were used. These dates coincide with the 2012 – 2016 drought in Southern California and were used to compare pre-drought, mid-drought, and post-drought conditions. I used Google Earth imagery to identify and draw polygons around 33 conifer stands within my study area. I then built a spectral library and used Multiple Endmember Spectral Mixture Analysis (MESMA) to compute sub-pixel fractional cover of green vegetation, non-photosynthetic vegetation, and soil for each of the three years. The years were then compared to identify trends in vegetation fractional cover over the drought in the conifer stand polygons. Finally, a Digital Elevation Model (DEM) was used to identify patterns of environmental factors (such as aspect) that may have contributed to conifer mortality. Results indicate that green vegetation cover in conifer stands decreased from 78.5% in 2013 to 66.1% in 2019.

From 2012 to 2016, Southern California experienced its worst drought in the last 1200 years (Lund et al., 2018). Over the last two years, plant species have had the opportunity to recover from this drought. In this project, we hope to gain insight into how plant species distributions change during and after drought. We also hope to determine the relationship between Land Surface Temperature (LST) and individual plant species in 2018. In this project, we used Airborne Visible / Infrared Imaging Spectrometer (AVIRIS) data to generate a plant species map from 2018 imagery. These were combined with previously created plant maps in June 2013 and 2015 (Meerdink, 2018). In this project, we use the MODIS/ASTER Airborne Simulator (MASTER) to examine the relationship between green vegetation (GV) fraction and LST. In order to map plant species, it was necessary to create and optimize a spectral library of the different species present in the reserve. Using Multiple Endmember Spectral Mixture Analysis (MESMA) data, we generated both a species map and a GV fractional cover map for 2018 in Sedgwick. In addition, we used MESMA and plant maps to compare relative temperatures between species. We determined a large increase in GV from 2015 to 2018, which means the Sedgwick Reserve area very quickly rebounded after the drought. In addition, we used the GV fractional cover map to confirm the negative linear relationship between GV and LST. To determine if the same relationship could be depicted with NDVI and LST, we plotted the NDVI values versus LST in Sedgwick Reserve. We concluded that this negative correlation was best displayed by the NDVI versus LST plot. On each of the three days, Mediterranean grasses and agriculture displayed temperatures at least 10 Kelvin higher than California Live Oak and Gray Pine.

The Central Valley is California’s most agriculturally productive region, where nearly 40% of water for irrigation is supplied by groundwater. During the 2012-2016 drought, surface water delivery to the region decreased by nearly half, leading to increased extraction of groundwater and a faster rate of ground subsidence (Tortajada et al., 2017). Building on the work of Shivers et al. 2018, this study uses a Random Forest algorithm and county-level crop data grouped according to California Department of Water Resources (DWR) guidelines to classify imagery from the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS). Random Forest is run based on a 50% minimum green vegetation fraction from AVIRIS as estimated by a multiple-endmember spectral mixture analysis (MESMA). Pixel classification accuracy exceeded 75% on average. According to Shivers et al., crop planting decisions in 2013-2015 during the drought were likely based on crop value and permanence rather than water usage. We expect to see a continuation of these findings with little to no evidence of a transition back to annual, lower-value crops from 2015-2018. Furthermore, drought intensity has been strongly correlated with ground subsidence, which can contribute to permanent loss of groundwater storage, reduced soil porosity, and the release of trapped mobile arsenic. These factors have the potential to influence crop planting decisions in the Central Valley based on rooting depth and differential uptake of contaminants. By using Sentinel-1 synthetic aperture radar (SAR) to map changes in ground surface elevation at a field-level resolution, we hope to find a relationship with spatial and temporal crop distributions.

California has long been at the forefront of Solar Energy production in the United States and has reaped the economic and environmental benefits from over 61 billion dollars in investment in the industry. Incentives for residential and utility-scale solar development have resulted in solar providing over 19 percent of California’s electricity as of the first quarter in 2019, according to the Solar Energy Industries Association. Grid monitoring and operation is difficult with the growth of solar, as over 1300 firms install a significant proportion of the market in urban residential areas. The objective of this research project is to test the effectiveness of imaging spectrometry for reliably quantifying urban photovoltaic panels. AVIRIS Next Generation flight data products taken over Santa Barbara for 2014 and 2019 and Multiple Endmember Spectral Mixture Analysis was implemented to map sub-pixel fractions of urban materials including vegetation, panels and other impervious surfaces. The fractional maps of these materials were then clipped into American Census Survey block groups to analyze the connections between household income and patterns of growth in panel coverage over the five years. Lastly, in 2018, Stanford’s “DeepSolar” Project received much acclaim for effectively mapping over 1.5 million panels in the United States through “machine learning” techniques applied on over a billion satellite images. Their Census tract level data has also been used in this project to compare panel coverage between the two methods in terms of quantity and spatial accuracy. The intended results from the methods I use could be extrapolated onto the larger “Deep Solar” study area if a combination of the two can provide the most optimal solar distribution map for the United States.

Biodiversity is important to maintain a healthy ecosystem and for the ecosystem to respond to disturbances with resilience. In Santa Barbara, there has been a visible shift in the distribution of plant species. Mapping the distribution of native and non-native species can show distribution changes and help with future predictions for species management. In this study, we mapped changes in the distribution of non-native and invasive species in Santa Barbara from 2013 to 2019. We utilized plant species polygons for locations of different species in Santa Barbara from 2013-2015 (Meerdink, 2018). These polygons were overlaid on 2019 imagery for comparison. Spectral libraries were developed and pruned from these polygons to identify the optimal spectra for mapping plant species. We randomly sampled from the spectra to display classifications of the species in maps. The study area and images were taken from Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) in 2013 and 2015 and AVIRIS Next Generation in 2019. Using a 2-endmember model, we ran Multiple Endmember Spectral Mixture Analysis (MESMA) within VIPER Tools to classify vegetation in Santa Barbara, CA, in 2019, 2015, and 2013. Species classifications derived from MESMA were binned into native vs. non-native and soil and urban areas were classified as non-vegetated areas. There were discrepancies in species and distributions between years because of changes in land use and phenomena such as fires and droughts in the area. Since MESMA was run with a 2-endmember model, areas with that had high amounts of mixed pixels became unmapped and were not shown in the final products. Changes between AVIRIS Classic imagery in 2013 and 2015 with AVIRIS 2019 NG data showed discrepancies between the accuracy of the species location. Some areas showed minimal change in species distribution, but other areas had larger variability in shifts between native and non-native.

Metrosideros polymorpha, also known as ʻŌhiʻa Lehua, is an endemic tree to Hawai’i that is very crucial for the preservation of the native ecosystems and watersheds. It is estimated that ʻōhiʻa trees make up 50 percent of all forests in Hawai’i and 80 percent of all native forests located in the state. A foreign fungal pathogen, known as rapid ʻōhiʻa death (ROD) was introduced to Hawai’i Island in 2010. ROD is known to turn the tree’s leaves brown and kill it in the span of a couple months. The goal of this project is to develop an index that locates areas of ʻōhiʻa that are infected with ROD in order to visualize possible trends of ROD to mitigate the pathogen. There is a noticeable difference in the spectral signature between the brown leaves of a ROD infected ʻōhiʻa and a green ʻōhiʻa. Through analysis of the spectral differences between infected and live ʻōhiʻa, an index called the Rapid Ohia Death Index (RODI) is developed. RODI can be used with the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) data to detect ʻōhiʻa trees that are infected with ROD. AVIRIS data from the years 2007, 2017, and 2018 is used to track the spatial and temporal spread of ROD in Ola’a Reserve, an ʻōhiʻa forest located just northeast of Volcano, Hawaii. Since most ROD detection programs are either on the ground or observing through a helicopter, survey areas are small and costly. Using RODI, larger areas can be surveyed to locate and mitigate ROD infected ʻōhiʻa to prevent further spreading of the pathogen. Results show obvious ROD infection in the forest from the 2017 data and a large increase in the amount of ROD infected ʻōhiʻa in 2018. RODI did not show a significant amount of ROD infections for the 2007 data, further validating the index because the pathogen was introduced to Hawai’i Island in 2010. The spatial spread of ROD seems to be moving westward, which follows the direction in which the east tradewinds blow.

Land surface temperature (LST) affects the temperature we feel, our productivity, and our health. Studies have shown that an increase in temperature raises morbidity and mortality rates (Deschenes and Greenstone 2011). In this project, we investigate how LST in Los Angeles varies based on income and evaluate how distance from the coast, fractional vegetation cover, and elevation affect the relationship between LST and income. For LST measurements, we used MODIS/ASTER (MASTER) thermal imagery from the HyspIRI Airborne Campign between 2013 and 2018. We examined 12 flight dates of the same flight line, which covers a range of socioeconomic properties and land cover. For evaluating elevation of the area, we used the ASTER Global Digital elevation model. To determine fractional vegetation cover, we ran LandSat 8 OLI images through Multiple Endmember Spectral Mixture Analysis (MESMA). We used the spectral library from Wetherley et al. (2018) to find the tree, turf, soil, roof, paved, non-photosynthetic vegetation (NPV), and shade fractions in LA County, then combined tree and turf fractions to find fractional vegetation cover. After preprocessing the data, we mapped mean LST, fractional vegetation cover, elevation, and coastal distance per US census tract for each flight. We found as median household income increased, LST decreased. For each flight, we created a linear regression model relating fractional vegetation cover, elevation, and coastal distance to LST. Though each date’s model differed, generally, fractional vegetation cover was the dominant factor, followed by coastal distance then elevation.

Accurate interpretation of phytoplankton functional types (PFTs) is important for monitoring harmful algal blooms and assuring public safety in aquatic environments. Previous work has suggested that it is possible to determine PFTs using remote sensing techniques on hyperspectral imagery such as Airborne Visible/InfraRed Imaging Spectrometer (AVIRIS) data. Here, we evaluate two different algorithms applied to the same remotely-sensed images to determine PFT proportions within the images. PHYtoplankton Detection with Optics Taxonomy (PHYDOTax, Palacios et al., 2015) is an algorithm written specifically for determining PFTs. However, its simplicity comes at the cost of high uncertainty depending on the complexity of the environment in the image. In contrast, Multiple Endmember Spectral Mixture Analysis (MESMA, Roberts et al., 2019) is a spectral unmixing technique primarily used to classify terrestrial remote sensing images down to the sub-pixel scale. The San Francisco Bay salt ponds were chosen as the study site to compare the two algorithms, given the unique aquatic environment containing many different PFTs, as well as the access to in-situ data to ground-truth the results from both algorithms. The validation of a reliable and accurate algorithm for determining PFT proportions will provide greater confidence when using remote sensing data to measure phytoplankton composition. Validated use of MESMA will also allow for simultaneous determination of other non-phytoplankton endmembers important for coastal waters, such as colored dissolved organic matter and total suspended sediment, in future applications.

Monitoring phytoplankton in coastal regions using remote sensing validation for satellite data allows scientists to more easily mitigate health hazards resulting from phytoplankton. Phytoplankton are microorganisms that serve as the base of the food web in aquatic ecosystems. Their behavior is difficult to predict since they are dynamic and incredibly variable as a group. For example, some produce toxins while the majority do not. Although many phytoplankton do not form chains under any circumstances, some phytoplankton form chains in low-stress environments. Phytoplankton respond to changes in nutrients. Upwelling, land interactions, and oceanic conditions in coastal waters influence nutrient and phytoplankton levels. Monterey Bay is an interesting place to study phytoplankton since these conditions shift, creating varied environments for phytoplankton life. The Kudela Lab located in Monterey Bay gathers in situ data from an instrument called the Imaging Flow Cytobot (IFCB) which collects images of phytoplankton in order to understand how phytoplankton populations changes in response to the environment. Using 2015 data from the IFCB, this project trained an existing machine learning algorithm to differentiate between phytoplankton that do and do not produce chains. This classifier performed with an overall error rate of 21%. A time series was constructed with the data showing the amount of chained phytoplankton in Monterey Bay. We looked for correlations to toxin levels (domoic acid) and found a small but potentially significant relationship. We also looked for spectral changes from MODIS Aqua between days with high numbers of chains versus days with no chains. Although we expected the amount of chains in the water to change the back scattering and thus the remote sensing reflectance, we did not see a significant change in MODIS spectra. Future projects would include lowering the classifier error and analyzing higher resolution spectral data against the chained phytoplankton.

Giant Kelp (Macrocystis pyrifera) is a macro algae that helps produce the oxygen marine organisms need to survive, and is considered a foundation species that provides habitat for many other organisms. Kelp forests are vital for maintaining a productive ecosystem that supports a variety of marine life in the Santa Barbara Channel, and in many other coastal ecosystems. Keeping the pH of the ocean slightly basic is imperative to maintaining the health of its inhabitants. The excess carbon dioxide in the atmosphere is causing the ocean to become more acidic due to the process of oceanic-atmospheric mixing. Giant kelp absorbs dissolved carbon dioxide in the ocean through photosynthesis, reducing the amount of carbon dioxide in the ocean. This project aims to determine if kelp forests are able to mitigate the progression of ocean acidification at a local scale. We examined the effect of kelp biomass and physiological health on the pH of the ocean near Isla Vista point in the Santa Barbara Channel. An algorithm previously developed by Cavanaugh et al. (2010) was used to calculate kelp biomass using Landsat-8 images. Oceanic pH data received from Santa Barbara Long Term Ecological Research and the National Oceanic and Atmospheric Administration were then compared with the kelp biomass data. We expect that the pH of the ocean near Isla Vista would have a positive correlation with the density of kelp in the area, supporting the proposed use of kelp as a “blue carbon” mechanism to reduce carbon dioxide levels in the ocean.

The coastal ocean is a dynamic environment which is constantly being influenced by many different factors. The 2011-2017 drought was one of the most severe in California history, but how this drought affected the coastal environment is not well understood. After the major drought, sediment and nutrients were washed into the Santa Barbara Channel (SBC) which prompted other attributes of the water to react in order to maintain balance. Using images from MODIS Aqua (Moderate Resolution Imaging Spectroradiometer) and Santa Barbara LTER (Long-Term Ecological Research) data, comparisons between sediment, pH, and chlorophyll were made to determine how the ocean responded to the added sediment following the drought. Sediment in the coastal ocean plays a dual role because runoff from land brings nutrients which help ocean life to thrive, but too much sediment can limit light availability and prevent phytoplankton from photosynthesizing. Sediment loads can also be a proxy for contaminants and other compounds which may hinder, be neutral, or promote ocean life. During the most severe years of the drought, runoff decreased substantially, but large rain events in the 2016-17 water year brought large quantities of sediment into the SBC. During this period of increased runoff, sediment plumes were easily detected in the Channel, and as the sediment concentration increased, pH decreased and chlorophyll increased. However this small time period might not be characteristic of the drought, so the analysis of chlorophyll, pH and sediment was extended to the whole drought period to quantify any long-term trends. Understanding how these three attributes affect one another can help us to understand how this complex environment changes over time and how humans influence it.

Lake Elsinore is Southern California’s largest freshwater lake. The economy of the surrounding community depends on revenue generated by the lake’s tourism industry. However, Lake Elsinore is prone to toxic algal blooms of Microcystis, a cyanobacteria that produces microcystin toxins. This project aims to determine how the toxins are aerosolized by lake activity and where the toxins go once airborne in the atmosphere. Previous research has shown that microcystins do aerosolize from water to the atmosphere. The natural water level flux of the lake was altered when the county authority constructed water/oxygen level regulators to ensure a stable ecosystem. Given that the lake can no longer filter away its nutrients during the dry season, nutrients accumulate and create a prime growing environment for algae to bloom. In this study, three sites around the perimeter of the lake were sampled for aerosolized microcystins, water quality and apparent optical properties. All sites had low levels of toxins. Site 1 had a reading of 2.5 ppb of microcystin in the water sample, while the air sample detected trace amounts of microcystin. Time series analysis was completed on the three sample sites as well as for the entire lake using Sentinel-2 imagery. Based on time series of fluorescence line height (FLH), the three sites show correlations in bloom frequency. NOAA HYSPLIT modeling was used to determine the relative movement of the wind over and around the lake. In the winter the air blew in the Southwest direction while in the other seasons the wind blew generally to the east. FLH values serve as proxies to relative chlorophyll values. Surrounding areas of highest exposure likelihood can be determined by combining HYSPLIT models and FLH algorithms.

It is important to keep track of whale locations because of their vulnerability to human and ecosystem perturbations. Whale populations of all types sunk to their lowest known levels in the 20th century after extensive whaling around the world. While global whale populations have been slowly recovering through adoption of regulations, it is necessary to continue to track whales to understand how populations move and grow, and under what circumstances they are likely to negatively encounter humans. In a 2014 paper, Fretwell et al. attempted to track southern right whales via satellite imagery using WorldView-2, which has eight color bands and a pixel size of 50 centimeters. They found that not only was it possible to track whales at the surface of the water based on their reflectance, but also to track whales below the surface. In this project, I used AVIRIS-NG hyperspectral remote sensing to try and track humpback whales off the coast of California. AVIRIS-NG has better spectral resolution with 450 color bands, but much lower spatial resolution at about 5 meters. In light of Fretwell’s paper, the question here was whether it was possible and practical to track whales in a similar manner using an airborne sensor. The first step was to simulate the eight color bands that WorldView-2 uses to directly compare to AVIRIS-NG. The reflectance signature of suspect whales was then determined in AVIRIS-NG to create a more accurate reflectance spectrum. The addition of more color bands in AVIRIS-NG leads to a more nuanced spectrum, which may make tracking whales more feasible, even though the spatial resolution is lower.

The ocean is an expansive but not limitless habitat accounting for 71% of the Earth’s surface. Plastic pollution in the ocean is a growing ecological concern for pelagic ecosystems because of its slow degradation time and ability to absorb and concentrate pollutants. Investigations into locating ocean plastic pollution have most commonly relied on limited in-situ collection and modeling, but have been limited by their lack of quantitative data. Ocean remote sensing of plastic pollution has been attempted with uncertain results due to (1) the lack of supportive in-situ data and (2) the limited spectral and spatial range for plastic at the ocean’s surface. Garaba et al. (2018) found that utilizing peaks in the visible and near infrared (NIR) appear promising for detecting plastic in hyperspectral data. This investigation examined four plastic types—ghost net, lids, bottles, and bags—comprised of varying types of polyethylene in a preliminary study by obtaining visible and NIR optical data of these plastics at varying depths in a swimming pool using a handheld spectral radiometer (ASD Fieldspec Handheld-2). Reflectance analysis revealed plastic signal attenuates exponentially with depth and disappears after 50cm. As most ocean plastic resides in the upper 50cm of the water column, it is possible that satellite imagery can discern plastic signals beyond the ocean’s surface. This study examined the feasibility and also provides the framework for future algorithms to detect ocean plastic through remote sensing.

The San Joaquin Valley (SJV) of California is home to approximately 4.2 million residents with an expected population increase of 60% within the next four decades.  Due to the exceptional levels of anthropogenically-induced combustion and agricultural activity occurring in the California Central Valley, this increase raises concern over air quality planning and controls in this region. Examining the seasonal and spatial variability of NO, NO₂, NO₃^–, and NH₄⁺, along with understanding the correlations between these compounds, can enhance our understanding of air quality over the SJV.  Therefore, this project focuses on the analysis of datasets compiled across three different airborne research flight campaigns: the 2010 CalNex and 2016 SARP summer campaigns, and the 2013 winter DISCOVER-AQ campaign. Differences in the observed seasonality of pollutants across each campaign provides key implications for understanding the temperature dependence involved in the productivity of gas-to-particle conversion reactions. The relative humidity dependence on the efficiency of these reactions is also considered. High resolution bins (0.5° × 0.5°) are mapped within SJV, including major cities such as Fresno and Bakersfield, to determine the spatial variability of the gases and aerosols of interest across all three campaigns. Results indicate that lower temperatures combined with higher relative humidities serve as two key parameters in enhancing the effectiveness of gas-to-particle conversions of the nitrogen-containing inorganic components.

In June 2016, the Sherpa Fire blazed through Santa Barbara County, California, burning over 7,000 acres of land in both the Santa Barbara Region and Los Padres National Forest. The transport of emissions from the Sherpa Fire was influenced by the Catalina Eddy, a typical meteorological phenomenon observed off the coast of Southern California from late spring to early fall. On June 17, 2016, the NASA DC-8 conducted 11 missed approaches at airports throughout the Los Angeles basin as part of the KORUS-AQ campaign. The DC-8 sampled the Sherpa Fire plume at different altitudes during these missed approaches, providing an in-depth look at the spatiotemporal variability of the plume due to transport by the Catalina Eddy. This study aims to track the transport of the Sherpa Fire plume by analyzing vertical profiles of acetonitrile in the LA basin. Acetonitrile, which is an excellent tracer for incomplete biomass combustion, was found to be correlated with aerosol concentration in smoke-influenced air masses. Wind data collected from multiple Automated Surface Observing System (ASOS) stations were used to evaluate the influence that the Catalina Eddy had on the transport of the Sherpa Fire plume throughout the LA basin.

Secondary organic aerosols (SOAs) are formed in the atmosphere from oxidation reactions involving volatile organic compounds (VOCs). SOAs are a major component of fine particulate matter, deteriorating air quality. Additionally, they contribute to cloud formation and the scattering or absorption of radiation, having an important impact on climate. SOA formation yields from specific VOCs have been measured in the laboratory and reported in literature. SOA formation potential (SFP), defined as the product of the specific VOC concentration and its SOA formation yield, indicates how much SOA can be produced from a given amount of VOC in the atmosphere. Using SOA yield data from Shin et al. (2012) and airborne Whole Air Sampling (WAS) data from ARCTAS 2008, FRAPPÉ 2014, and SARP 2012 and 2016, the regional and annual variability of SFP were investigated. The analysis allowed a comparison among SFPs in different regions, namely Colorado, the Central Valley, LA Basin, and Antelope Valley. Furthermore, data from each summer was averaged and compared to investigate SFP variability in California from 2008 to 2016. This study found that during summer in California and Colorado, benzene was the highest contributor to SOA formation, followed by toluene. Total SFP was lowest in 2012 and highest in 2008.

Over the last two centuries, anthropogenic pollution has altered the planet’s radiative balance; however, quantifying this alteration is difficult because of Earth’s complex interconnected systems. Interactions between aerosols and clouds are the largest source of uncertainty in quantifying the anthropogenic impact on climate. Clouds can either warm or cool the planet depending on several factors like thickness and height in the sky. To determine the extent of clouds’ impact on the Earth’s radiative balance, one must consider the potential for cloud formation through cloud condensation nuclei (CCN), which are the number of particles in the atmosphere capable of forming cloud droplets. This work investigates the spatial distribution of CCN activity in California’s Central Valley during the DISCOVER-AQ 2013 campaign in winter and the SARP 2019 campaign in summer. To visualize CCN activity, the ratio of CCN:CN (where CN is all particles greater than 10 nm) was mapped in 0.5 degree latitude and longitude pixels. When comparing the two datasets, it was found that CCN:CN was overall lower in the summer than the winter. The DISCOVER-AQ dataset allowed investigating aerosol composition and size distributions in conjunction with CCN:CN. This analysis suggested that bulk inorganic composition and size did not appear to be influencing the spatial distribution of CCN activity. During both missions, the nitrogen oxide ratio, NO_x:NO_y, was also mapped as a proxy for air mass age. In winter, NO_x:NO_y was strongly anticorrelated with CCN:CN meaning that as the aerosol aged it became more CCN active. During the summer, this correlation deteriorated. Further work will focus on characterizing aerosol hygroscopic properties during the DISCOVER-AQ campaign.

Aerosols can alter the Earth’s incoming and outgoing radiation by either absorbing or scattering radiation. Since the global radiation budget is in constant flux, understanding more about the influence of aerosols in this budget is needed. Depending on their chemical and hence optical properties, aerosols can either reflect or absorb solar radiation, causing either a cooling or heating effect, respectively. While aerosols generally provide a cooling effect, the extent of cooling depends on aerosol composition, optical properties, and size. Using flights from the 2016 NASA Student Airborne Research Program (SARP), measurements of chemical composition from the Aerosol Mass Spectrometer (AMS), size distribution from the Laser Aerosol Spectrometer (LAS), and Single Scattering Albedo (SSA) from the Langley Aerosol Research Group Experiment (LARGE) were used to characterize the radiative forcing efficiency of aerosols. To do so, calculations were performed in Python using Mie theory. The study region included the Inland Empire, the LA Basin, and the Central Valley, which have distinct air mass types. Regional variations of radiative forcing were found primarily due to differences in SSA values and particle diameter. Further exploration is needed to distinguish between the relative influence of SSA and size, as well as incorporating variability in surface reflectance between the aforementioned study regions.

Biogeochemical cycles govern the movement and distribution of nutrients, the variety of compounds crucial for life. Varying amounts of these components can then determine ecosystem growth and change. This study focuses on potassium, iron, and phosphorus, as well as wildfires, which are able to emit and transport these nutrients to locations away from the source. For this study, California data measured in 2016 from the Interagency Monitoring of Protected Visual Environments (IMPROVE) system was used to track if and when nutrients appear at the surface from biomass burning events. Satellite imagery and the Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model was used to validate when peaks in nutrient values occur spatially and temporally. Nutrient-to-black carbon emission ratios were found for the Sherpa Fire using ground data. Combining these values with Student Airborne Research Program (SARP) 2016 data allows for an estimate of nutrient levels at high elevations. Approximate fluxes of dry aerosol deposition were also calculated for periods of time with nearby burning and periods without. To look at broad scale correlations, average concentrations for all active monitoring sites in each year were compared to the corresponding number of acres burned in the state of California. Overall, the data is mixed but shows some clear indicators that fire emissions influence the distribution and deposition of these nutrients. Further studies can utilize data from additional years in addition to oceanic and precipitation measurements to gain a better understanding of the impacts of fires.

The Salton Sea is a large endorheic lake that is drying, thus exposing the lakebed that was once submerged. This lakebed, known as playa, emits both gases and particles into the atmosphere that have adverse effects on local air quality. Aircraft data from the 2019 Student Airborne Research Program (SARP) were used in order to determine the vertical distribution of aerosols over the Salton Sea to determine the overall effect of these emissions on the atmosphere. During a high-altitude pass at 3,200 meters over the Salton Sea, the Differential Absorption LIDAR (DIAL) instrument was used in order to analyze the remotely sensed aerosol optical properties. Following this, a lower altitude pass at 320 meters provided an opportunity to confirm DIAL data through comparison with the in-situ Langley Aerosol Research Group Experiment (LARGE) instruments as well as the NOAA In-Situ Measurements of Aerosol Optical Properties (NOAA-AOP) instrument. In order to differentiate between aerosol types, we referenced a published algorithm which creates eight classes of aerosol based on the lidar ratio at 532nm, backscatter color ratio, aerosol depolarization spectral ratio, and aerosol depolarization at 532nm. By further dividing the Salton Sea into three geographic regions, we analyzed the different aerosol types that make up the vertical layers in the lowest kilometer of the atmosphere, and identified which of the three distinct vertical layers were emitted from the Salton Sea and surrounding playa. The lowest 500 meters of the atmosphere in the region was shown to be impacted by the Salton Sea, as aerosol at this elevation shared similar optical properties to maritime aerosol.