Hyperspectral unmixing is an important remote sensing task with applications including material identification and analysis. Characteristic spectral features make many pure materials identifiable from their visible-to-infrared spectra, but quantifying their presence within a mixture is a challenging task due to nonlinearities and factors of variation. In this thesis, physics-based approaches are incorporated into an end-to-end spectral unmixing algorithm via differentiable programming. First, sparse regularization and constraints are implemented by adding differentiable penalty terms to a cost function to avoid unrealistic predictions. Secondly, a physics-based dispersion model is introduced to simulate realistic spectral variation, and an efficient method to fit the parameters is presented. Then, this dispersion model is utilized as a generative model within an analysis-by-synthesis spectral unmixing algorithm. Further, a technique for inverse rendering using a convolutional neural network to predict parameters of the generative model is introduced to enhance performance and speed when training data are available. Results achieve state-of-the-art on both infrared and visible-to-near-infrared (VNIR) datasets as compared to baselines, and show promise for the synergy between physics-based models and deep learning in hyperspectral unmixing in the future.

The 78 secondary eclipse depths for a sample of 36 transiting hot Jupiters observed at 3.6- and 4.5 μm using the Spitzer Space Telescope is here reported. Eclipse results for 27 of these planets are new and include highly irradiated worlds such as KELT-7b (Kilodegree Extremely Little Telescope), WASP-87b (Wide Angle Search for Planets), WASP-76b, and WASP-64b, and important targets for the James Webb Space Telescope (JWST) such as WASP-62b. WASP-62b is found to have a slightly eccentric orbit (ecosω=0.00614±0.00058), and the eccentricities of HAT-P-13b (Hungarian Automated Telescope Project) and WASP-14b are confirmed. The remainder are individually consistent with circular orbits, but there is statistical evidence for eccentricity increasing with orbital period in this range from 1 to 5 days. Day-side brightness temperatures (Tb) for the planets yield information on albedo and heat redistribution, following Cowan and Agol (2011). Planets having maximum day side temperatures exceeding ∼2200 K are consistent with zero albedo and distribution of stellar irradiance uniformly over the day-side hemisphere. The most intriguing result is a detection of a systematic difference between the emergent spectra of these hot Jupiters as compared to blackbodies. The ratio of observed brightness temperatures, Tb(4.5)/Tb(3.6), increases with equilibrium temperature by 98±26 parts-per-million per Kelvin, over the entire temperature range in the sample (800K to 2500K). No existing model predicts this trend over such a large range of temperature. This may be due to a structural difference in the atmospheric temperature profile between the real planetary atmospheres as compared to models.

Recurring Slope Lineae (RSL) are dark, narrow features which form on steep Martian slopes during warm seasons, lengthening, fade in cold seasons and recurring annually. There are many hypotheses on the formation mechanism of RSL. A number of these hypotheses suggest that RSL are liquid brines flowing on the surface. Brine based hypotheses often state that sub-surface aquifers are necessary to supply the water needed to recharge RSL. One problem with this is that RSL are observed forming on isolated peaks and ridgelines where a sub-surface aquifer is unlikely. This study uses a thermal model called KRC to examine the correlation between RSL activity and surface temperature at several RSL sites in Valles Marineris. This correlation is compared to the freezing temperature of several brines. Results show an interesting relationship between RSL activity and the surface temperature of very steep (> 60º) slopes. This could indicate that RSL are caused by thermal stresses loosening material on the face of bedrock outcroppings instead of briny flows.

Emission Spectroscopy is a powerful tool for the identification of mineralogical samples and has been used for decades in labs to study the geology of Earth and Mars. However, the instruments needed to make these measurements are large, expensive and sensitive pieces of equipment that are too cumbersome to use in the field. There are some commercial products that attempt to work in the field, however they perform this task poorly, often resulting in limited applications, poor performance or not being truly portable. My thesis utilizes the TES family of planetary instruments as a source of inspiration for creating a truly portable Fourier Transform InfraRed spectrometer. From this initial design phase, it appears that it is possible to build an instrument with vastly improved capabilities over the current systems on the market. This roughly 12 inch by 7 inch by 8 inch device with a 3-inch diameter telescope is capable of achieving a SNR of over 1000 during a 5 minute scan of a sample allowing for 5 sigma (99.99994% Confidence) identification of 1% spectral features from 5 um to >60 um making this instrument a one of a kind device with high application potential, not only for field geologist but for the future of manned exploration of space. Currently an accurate measurement of costs is not available, however with more development and optimization a total cost of around $50K is feasible while still maintaining the same performance characteristics. If the costs can fall within an acceptable range, this device will not only be technically impressible but viable from a financial standpoint as well.

Large, violent storms come through the Phoenix area during monsoon season, and currently, the best ways to predict them are not very accurate. The primary goal of this investigation is to see if a mechanism can be developed for the prediction of these storms in Phoenix during monsoon season. In order to answer this question, two data sets (a remote sensing satellite imagery and a ground-based weather information set) will be used and their measurements will be compared against one another using a corresponding time as the related variable. The goal is to try and identify some type of correlation or explanation of correlation. Events known as moisture surges (from the gulf surge \u2014 which comes from the California Gulf) will be identified and then compared in some detail. These chutes of moisture surge through Arizona, primarily up through Yuma in a northeasterly direction. The point of the investigation is to prove or disprove that satellite imagery can be used as an analog for dew point measurements in areas where ground measurements are not available. If this can be demonstrated, then, because of the high temporal resolution of the remote sensing data, satellite imagery could be used as an identifier of oncoming storms.