The marked increase in the inflow of remotely sensed data from satellites have trans- formed the Earth and Space Sciences to a data rich domain creating a rich repository for domain experts to analyze. These observations shed light on a diverse array of disciplines ranging from monitoring Earth system components to planetary explo- ration by highlighting the expected trend and patterns in the data. However, the complexity of these patterns from local to global scales, coupled with the volume of this ever-growing repository necessitates advanced techniques to sequentially process the datasets to determine the underlying trends. Such techniques essentially model the observations to learn characteristic parameters of data-generating processes and highlight anomalous planetary surface observations to help domain scientists for making informed decisions. The primary challenge in defining such models arises due to the spatio-temporal variability of these processes.

This dissertation introduces models of multispectral satellite observations that sequentially learn the expected trend from the data by extracting salient features of planetary surface observations. The main objectives are to learn the temporal variability for modeling dynamic processes and to build representations of features of interest that is learned over the lifespan of an instrument. The estimated model parameters are then exploited in detecting anomalies due to changes in land surface reflectance as well as novelties in planetary surface landforms. A model switching approach is proposed that allows the selection of the best matched representation given the observations that is designed to account for rate of time-variability in land surface. The estimated parameters are exploited to design a change detector, analyze the separability of change events, and form an expert-guided representation of planetary landforms for prioritizing the retrieval of scientifically relevant observations with both onboard and post-downlink applications.

Water has shaped the surface of Mars, recording previous environments and inspiring the search for extinct life beyond Earth. While conditions on the Martian surface today are not conducive to the presence of liquid water, ancient erosional and depositional features indicate that this was not always so. Quantifying the regional and global history of water on Mars is crucial to understanding how the planet evolved, where to focus future exploration, and implications for water on Earth.

Many sites on Mars contain layered sedimentary deposits, sinuous valleys with delta shaped deposits, and other indications of large lakes. The Hypanis deposit is a unique endmember in this set of locations as it appears to be the largest ancient river delta identified on the planet, and it appears to have no topographic boundary, implying deposition into a sea. I have used a variety of high-resolution remote sensing techniques and geologic mapping techniques to present a new model of past water activity in the region.

I gathered new orbital observations and computed thermal inertia, albedo, elevation, and spectral properties of the Hypanis deposit. I measured the strike and dip of deposit layers to interpret the sedimentary history. My results indicate that Hypanis was formed in a large calm lacustrine setting. My geomorphic mapping of the deposit and catchment indicates buried volatile-rich sediments erupted through the Chryse basin fill, and may be geological young or ongoing. Collectively, my results complement previous studies that propose a global paleoshoreline, and support interpretations that Mars had an ocean early in its history. Future missions to the Martian surface should consider Hypanis as a high-value sampling opportunity.

Jupiter’s moon Europa is an active target of research because of its unique geology and its potential for habitability. Europa’s icy chaos disrupts and transforms the previous terrain, suggesting melting is involved. Chaos occurs alongside several types of endogenic surface features. These microfeatures are under <100 km2 in area and include uplifts and domes, pits, spots, and hybrid features. The distribution of microfeatures is known in the ~10% of the Europa’s surface that are covered by the regional mosaics (“RegMaps”). The efforts to connect microfeature formation to any kind of heat transport in Europa are confounded because microfeatures are difficult to identify outside of RegMaps because of low image resolutions. Finding microfeatures outside of RegMaps would provide new observational constraints for microfeature formation models.

First, I mapped microfeatures across four of Europa’s RegMaps and validated them against other mapping datasets. Microchaos features are the most numerous, followed by pits, domes, then hybrids. Spots are the least common features, and the smallest. Next, I mapped features in low-resolution images that covered the E15RegMap01 area to determine error rates and sources of omission or misclassification for features mapped in low-resolution images. Of all features originally mapped in the RegMap, pits and domes were the least likely to be re-mapped or positively identified (24.2% and 5%, respectively). Chaos, spots, and hybrids were accurately classified over 70% of the time. Quantitatively classifying these features using discriminant function analysis yielded comparable values of accuracy when compared to a human mapper. Finally, nearest-neighbor clustering analyses were used to show that pits are clustered in all regions, while chaos, domes, and hybrids vary in terms of their spatial clustering.

This work suggests that the most likely processes for microfeature formations is either the evolution of liquid water sills within Europa’s ice shell or cryovolcanism. Future work extending to more areas outside of the RegMaps can further refine microfeature formation models. The detection of liquid water at or near the surface is a major goal of multiple upcoming Europa missions; this work provides predictions that can be directly tested by these missions to maximize their scientific return.

The Kilopixel Array Pathfinder Project (KAPPa) advances the number of coherent high-frequency terahertz (THz) receivers that could be packed into a single focal plane array on existing submm telescopes. The KAPPa receiver, at 655-695 GHz, is a high frequency heterodyne receiver that can achieve system temperatures of less than 200 K, the specification for ALMA band-9. The KAPPa receiver uses a novel design of a permanent magnet to suppress the noise generated by the DC Josephson effect. This is in stark contrast to the benchmark solution of an electromagnet that is both too expensive and too large for use in kilo-pixel arrays. I present a simple, robust design for a single receiver element that can be tessellated throughout a telescope's focal plane to make a ~1000 pixel array, which is much larger than the current state-of-the-art array, SuperCam, at 64 pixels and ~345 GHz.

While the original goal to develop receiver technologies has been accomplished, the path to this accomplishment required a far more holistic approach than originally anticipated. The goal of the present work has expended exponentially from that of KAPPas promised technical achievements. In the present work, KAPPa and its extension, I present solutions ranging from 1) the creation of large scale astronomical maps, 2) metaheuristic algorithms that solve tasks too complex for humans, and 3) detailed technical assembly of microscopic circuit components. Each part is equally integral for the realization of a ~1000 pixel THz arrays.

Our automated tuning algorithm, Alice, uses differential evolution techniques and has been extremely successful in its implementation. Alice provides good results for characterizing the extremely complex tuning topology of THz receivers. More importantly, it has accomplished rapid optimization of an entire array without human intervention. In the age of big data astronomy, I have prepared THz heterodyne receiver arrays by making cutting edge community-oriented data analysis tools for the future of large-scale discovery. I present a from-scratch reduction and analysis architecture developed for observations of 100s of square degree on-the-sky maps with SuperCam to address the gulf between observing with single dish antennas versus a truly integrated focal plane array.

ABSTRACT

The Spirit landing site in Gusev Crater has been imaged by the Mars Reconnaissance Orbiter High Resolution Imaging Science Experiment (HiRISE) camera more than thirty times since 2006. The breadth of this image set allowed a study of changes to surface features, covering four Mars years.

Small fields of bedforms comprised of dark material, and dark dust devil tracks are among the features revealed in the images. The bedforms are constrained within craters on the plains, and unconstrained in depressions less than 200m wide within the topography of the Columbia Hills, a ~120m-high structure in center of Gusev. Dust devil tracks appear in many images of the bedforms.

Within the Columbia Hills, three bedform fields approximately 180m2 and composed of fine dark basaltic sand were studied, using five HiRISE images taken from 2006 to 2014. Both bedform crests and the dust devil tracks superimposed on them were evaluated for change to azimuth and length, and for correlation between the features. The linear to slightly sinuous transverse crests ranging from less than 1m to 113m in length and two to three meters in wavelength, are primary bedforms. During the study they shifted as much as 33 degrees in azimuth, and individual crests moved on the surface as much as 0.75m. The greatest changes corresponded to a global dust storm in 2007. Average crest movement was documented at the rate of 0.25m per year. Rather than moving progressively, the crests eventually returned to near their original orientation after the storm. The dust devil tracks, reflecting a more complex wind regime, including vortex development during diurnal heating, maintained predominantly NW-SE orientations but also reflected the effects of the storm.

The observed modifications were neither progressive, nor strictly seasonal. The apparent stability of the bedform geometry over four seasons supports the predictions of the Mars Regional Atmospheric Modeling System (MRAMS): low speed (1-7.5 ms-1), daily alternating winds of relatively equal force. Crest profiles were found to be nearly symmetrical, without slipfaces to indicate a preferential wind direction; this finding also is supported by the MRAMS model.

Amorphous phases are detected over large regions of the Martian surface from orbit and in more localized deposits by rovers on the surface. Amorphous silicates can be primary or secondary in origin, both having formed through very different processes, so the unambiguous identification of these phases is important for understanding the geologic history of Mars. Secondary amorphous silicates are poorly understood and underrepresented in spectral libraries because they lack the long-range structural order that makes their crystalline counterparts identifiable in most analytical techniques. Fortunately, even amorphous materials have some degree of short-range order so that distinctions can be made with careful characterization.

Two sets of laboratory experiments were used to produce and characterize amorphous weathering products under probable conditions for the Martian surface, and one global spectral analysis using thermal-infrared (TIR) data from the Thermal Emission Spectrometer (TES) instrument was used to constrain variations in amorphous silicates across the Martian surface. The first set of experiments altered crystalline and glassy basalt samples in an open system under strong (pH 1) and moderate (pH 3) acidic conditions. The second set of experiments simulated a current-day Martian weathering scenario involving transient liquid water where basalt glass weathering solutions, formed in circumneutral (pH ~5.5 and 7) conditions, were rapidly evaporated, precipitating amorphous silicates. The samples were characterized using visible and near-infrared (VNIR) spectroscopy, TIR spectroscopy, scanning electron microscopy (SEM), energy dispersive X-ray spectroscopy (EDS), and X-ray diffraction (XRD).

All experiments formed amorphous silicate phases that are new to spectral libraries. Moderately acidic alteration experiments produced no visible or spectral evidence of alteration products, whereas exposure of basalt glass to strongly acidic fluids produced silica-rich alteration layers that are spectrally consistent with VNIR and TIR spectra from the circum-polar region of Mars, indicating this region has undergone acidic weathering. Circum-netural pH basalt weathering solution precipitates are consistent with amorphous materials measured by rovers in soil and rock surface samples in Gale and Gusev Craters, suggesting transient water interactions over the last 3 billion years. Global spectral analyses determine that alteration conditions have varied across the Martian surface, and that alteration has been long lasting.

Both volcanism and impact cratering produce ejecta and associated deposits incorporating a molten rock component. While the heat sources are different (exogenous vs. endogenous), the end results are landforms with similar morphologies including ponds and flows of impact melt and lava around the central crater. Ejecta from both impact and volcanic craters can also include a high percentage of melted rock. Using Lunar Reconnaissance Orbiter Camera Narrow Angle Camera (LROC NAC) images, crucial details of these landforms are finally revealed, suggesting a much more dynamic Moon than is generally appreciated. Impact melt ponds and flows at craters as small as several hundred meters in diameter provide empirical evidence of abundant melting during the impact cratering process (much more than was previously thought), and this melt is mobile on the lunar surface for a significant time before solidifying. Enhanced melt deposit occurrences in the lunar highlands (compared to the mare) suggest that porosity, target composition, and pre-existing topography influence melt production and distribution. Comparatively deep impact craters formed in young melt deposits connote a relatively rapid evolution of materials on the lunar surface. On the other end of the spectrum, volcanic eruptions have produced the vast, plains-style mare basalts. However, little was previously known about the details of small-area eruptions and proximal volcanic deposits due to a lack of resolution. High-resolution images reveal key insights into small volcanic cones (0.5-3 km in diameter) that resemble terrestrial cinder cones. The cones comprise inter-layered materials, spatter deposits, and lava flow breaches. The widespread occurrence of the cones in most nearside mare suggests that basaltic eruptions occur from multiple sources in each basin and/or that rootless eruptions are relatively common. Morphologies of small-area volcanic deposits indicate diversity in eruption behavior of lunar basaltic eruptions driven by magmatic volatiles. Finally, models of polar volatile behavior during impact-heating suggest that chemical alteration of minerals in the presence of liquid water is one possible outcome that was previously not thought possible on the Moon.

Among volcanic gases, sulfur dioxide (SO2) is by far the most commonly measured. More than a monitoring proxy for volcanic degassing, SO2 has the potential to alter climate patterns. Persistently active explosive volcanoes are characterized by short explosive bursts, which often occur at periodic intervals numerous times per day, spanning years to decades. SO2 emissions at those volcanoes are poorly constrained, in large part because the current satellite monitoring techniques are unable to detect or quantify plumes of low concentration in the troposphere. Eruption plumes also often show high concentrations of ash and/or aerosols, which further inhibit the detection methods. In this work I focus on quantifying volcanic gas emissions at persistently active explosive volcanoes and their variations over short timescales (minutes to hours), in order to document their contribution to natural SO2 flux as well as investigate the physical processes that control their behavior.

In order to make these measurements, I first develop and assemble a UV ground-based instrument, and validate it against an independently measured source of SO2 at a coal-burning power plant in Arizona. I establish a measurement protocol and demonstrate that the instrument measures SO2 fluxes with < 20 % error. Using the same protocol, I establish a record of the degassing patterns at Semeru volcano (Indonesia), a volcano that has been producing cycles of repeated explosions with periods of minutes to hours for the past several decades. Semeru produces an average of 21-71 tons of SO2 per day, amounting to a yearly output of 8-26 Mt.

Using the Semeru data, along with a 1-D transient numerical model of magma ascent, I test the validity of a model in which a viscous plug at the top of the conduit produces cycles of eruption and gas release. I find that it can be a valid hypothesis to explain the observed patterns of degassing at Semeru. Periodic behavior in such a system occurs for a very narrow range of conditions, for which the mass balance between magma flux and open-system gas escape repeatedly generates a viscous plug, pressurizes the magma beneath the plug, and then explosively disrupts it.

Much of Mars' surface is mantled by bright dust, which masks the spectral features used to interpret the mineralogy of the underlying bedrock. Despite the wealth of near-infrared (NIR) and thermal infrared data returned from orbiting spacecraft in recent decades, the detailed bedrock composition of approximately half of the martian surface remains relatively unknown due to dust cover. To address this issue, and to help gain a better understanding of the bedrock mineralogy in dusty regions, data from the Thermal Emission Spectrometer (TES) Dust Cover Index (DCI) and Mars Reconnaissance Orbiter (MRO) Mars Color Imager (MARCI) were used to identify 63 small localized areas within the classical bright dusty regions of Arabia Terra, Elysium Planitia, and Tharsis as potential "windows" through the dust; that is, areas where the dust cover is thin enough to permit infrared remote sensing of the underlying bedrock. The bedrock mineralogy of each candidate "window" was inferred using processed spectra from the Mars Express (MEx) Observatoire pour la Mineralogie, l'Eau, les Glaces et l'Activité (OMEGA) NIR spectrometer and, where possible, TES. 12 areas of interest returned spectra that are consistent with mineral species expected to be present at the regional scale, such as high- and low-calcium pyroxene, olivine, and iron-bearing glass. Distribution maps were created using previously defined index parameters for each species present within an area. High-quality TES spectra, if present within an area of interest, were deconvolved to estimate modal mineralogy and support NIR results. OMEGA data from Arabia Terra and Elysium Planitia are largely similar and indicate the presence of high-calcium pyroxene with significant contributions of glass and olivine, while TES data suggest an intermediate between the established southern highlands and Syrtis Major compositions. Limited data from Tharsis indicate low-calcium pyroxene mixed with lesser amounts of high-calcium pyroxene and perhaps glass. TES data from southern Tharsis correlate well with the previously inferred compositions of the Aonium and Mare Sirenum highlands immediately to the south.

In this thesis I model the thermal and structural evolution of Kuiper Belt Objects (KBOs) and explore their ability to retain undifferentiated crusts of rock and ice over geologic timescales. Previous calculations by Desch et al. (2009) predicted that initially homogenous KBOs comparable in size to Charon (R ~ 600 km) have surfaces too cold to permit the separation of rock and ice, and should always retain thick (~ 85 km) crusts, despite the partial differentiation of rock and ice inside the body. The retention of a thermally insulating, undifferentiated crust is favorable to the maintenance of subsurface liquid and potentially cryovolcanism on the KBO surface. A potential objection to these models is that the dense crust of rock and ice overlying an ice mantle represents a gravitationally unstable configuration that should overturn by Rayleigh-Taylor (RT) instabilities. I have calculated the growth rate of RT instabilities at the ice-crust interface, including the effect of rock on the viscosity. I have identified a critical ice viscosity for the instability to grow significantly over the age of the solar system. I have calculated the viscosity as a function of temperature for conditions relevant to marginal instability. I find that RT instabilities on a Charon-sized KBO require temperatures T > 143 K. Including this effect in thermal evolution models of KBOs, I find that the undifferentiated crust on KBOs is thinner than previously calculated, only ~ 50 km. While thinner, this crustal thickness is still significant, representing ~ 25% of the KBO mass, and helps to maintain subsurface liquid throughout most of the KBO's history.