arXiv subjects

Michael Mommert

Publications and source records attributed to Michael Mommert.

At least 37 records · Page 2Linked to original sources

arXiv2021

Power Plant Classification from Remote Imaging with Deep Learning

Satellite remote imaging enables the detailed study of land use patterns on a global scale. We investigate the possibility to improve the information content of traditional land use classification by identifying the nature of industrial sites from medium-resolution remote sensing images. In this work, we focus on classifying different types of power plants from Sentinel-2 imaging data. Using a ResNet-50 deep learning model, we are able to achieve a mean accuracy of 90.0% in distinguishing 10 different power plant types and a background class. Furthermore, we are able to identify the cooling mechanisms utilized in thermal power plants with a mean accuracy of 87.5%. Our results enable us to qualitatively investigate the energy mix from Sentinel-2 imaging data, and prove the feasibility to classify industrial sites on a global scale from freely available satellite imagery.

cs.CV↗

arXiv2020

Characterization of Industrial Smoke Plumes from Remote Sensing Data

The major driver of global warming has been identified as the anthropogenic release of greenhouse gas (GHG) emissions from industrial activities. The quantitative monitoring of these emissions is mandatory to fully understand their effect on the Earth's climate and to enforce emission regulations on a large scale. In this work, we investigate the possibility to detect and quantify industrial smoke plumes from globally and freely available multi-band image data from ESA's Sentinel-2 satellites. Using a modified ResNet-50, we can detect smoke plumes of different sizes with an accuracy of 94.3%. The model correctly ignores natural clouds and focuses on those imaging channels that are related to the spectral absorption from aerosols and water vapor, enabling the localization of smoke. We exploit this localization ability and train a U-Net segmentation model on a labeled sub-sample of our data, resulting in an Intersection-over-Union (IoU) metric of 0.608 and an overall accuracy for the detection of any smoke plume of 94.0%; on average, our model can reproduce the area covered by smoke in an image to within 5.6%. The performance of our model is mostly limited by occasional confusion with surface objects, the inability to identify semi-transparent smoke, and human limitations to properly identify smoke based on RGB-only images. Nevertheless, our results enable us to reliably detect and qualitatively estimate the level of smoke activity in order to monitor activity in industrial plants across the globe. Our data set and code base are publicly available.

cs.CV↗

arXiv2020

Distinguishing multicellular life on exoplanets by testing Earth as an exoplanet

Can multicellular life be distinguished from single cellular life on an exoplanet? We hypothesize that abundant upright photosynthetic multicellular life (trees) will cast shadows at high sun angles that will distinguish them from single cellular life and test this using Earth as an exoplanet. We first test the concept using Unmanned Arial Vehicles (UAVs) at a replica moon landing site near Flagstaff, Arizona and show trees have both a distinctive reflectance signature (red edge) and geometric signature (shadows at high sun angles) that can distinguish them from replica moon craters. Next, we calculate reflectance signatures for Earth at several phase angles with POLDER (Polarization and Directionality of Earth's reflectance) satellite directional reflectance measurements and then reduce Earth to a single pixel. We compare Earth to other planetary bodies (Mars, the Moon, Venus, and Uranus) and hypothesize that Earths directional reflectance will be between strongly backscattering rocky bodies with no weathering (like Mars and the Moon) and cloudy bodies with more isotropic scattering (like Venus and Uranus). Our modelling results put Earth in line with strongly backscattering Mars, while our empirical results put Earth in line with more isotropic scattering Venus. We identify potential weaknesses in both the modeled and empirical results and suggest additional steps to determine whether this technique could distinguish upright multicellular life on exoplanets.

astro-ph.EP↗

arXiv2020

Cloud Identification from All-sky Camera Data with Machine Learning

Most ground-based observatories are equipped with wide-angle all-sky cameras to monitor the night sky conditions. Such camera systems can be used to provide early warning of incoming clouds that can pose a danger to the telescope equipment through precipitation, as well as for sky quality monitoring. We investigate the use of different machine learning approaches for automating the identification of mostly opaque clouds in all-sky camera data as a cloud warning system. In a deep-learning approach, we train a Residual Neural Network (ResNet) on pre-labeled camera images. Our second approach extracts relevant and localized image features from camera images and uses these data to train a gradient-boosted tree-based model (lightGBM). We train both model approaches on a set of roughly 2,000 images taken by the all-sky camera located at Lowell Observatory's Discovery Channel Telescope, in which the presence of clouds has been labeled manually. The ResNet approach reaches an accuracy of 85% in detecting clouds in a given region of an image, but requires a significant amount of computing resources. Our lightGBM approach achieves an accuracy of 95% with a training sample of ~1,000 images and rather modest computing resources. Based on different performance metrics, we recommend the latter feature-based approach for automated cloud detection. Code that was built for this work is available online.

astro-ph.IM↗

arXiv2019

The Luminosity Function of Red Supergiants in M31

The mass-loss rates of red supergiant stars (RSGs) are poorly constrained by direct measurements, and yet the subsequent evolution of these stars depends critically on how much mass is lost during the RSG phase. In 2012 the Geneva evolutionary group updated their mass-loss prescription for RSGs with the result that a 20 solar mass star now loses 10x more mass during the RSG phase than in the older models. Thus, higher mass RSGs evolve back through a second yellow supergiant phase rather than exploding as Type II-P supernovae, in accord with recent observations (the so-called "RSG Problem"). Still, even much larger mass-loss rates during the RSG phase cannot be ruled out by direct measurements of their current dust-production rates, as such mass-loss is episodic. Here we test the models by deriving a luminosity function for RSGs in the nearby spiral galaxy M31 which is sensitive to the total mass loss during the RSG phase. We carefully separate RSGs from asymptotic giant branch stars in the color-magnitude diagram following the recent method exploited by Yang and collaborators in their Small Magellanic Cloud studies. Comparing our resulting luminosity function to that predicted by the evolutionary models shows that the new prescription for RSG mass-loss does an excellent job of matching the observations, and we can readily rule out significantly larger values.

astro-ph.SR↗

arXiv2019

Visible spectroscopy from the Mission Accessible Near-Earth Object Survey (MANOS): Taxonomic dependence on asteroid size

The Mission Accessible Near-Earth Object Survey (MANOS) aims to observe and characterize small (mean absolute magnitude H ~ 25 mag) Near-Earth Objects (NEOs) that are accessible by spacecraft (mean $Δv$ ~ 5.7 km/s) and that make close approaches with the Earth (mean Minimum Orbital Intersection Distance MOID ~ 0.03 AU). We present here the first results of the MANOS visible spectroscopic survey. The spectra were obtained from August 2013 to March 2018 at Lowell Observatory's Discovery Channel 4.3 meter telescope, and both Gemini North and South facilities. In total, 210 NEOs have been observed and taxonomically classified. Our taxonomic distribution shows significant variations with respect to surveys of larger objects. We suspect these to be due to a dependence of Main Belt source regions on object size. Compared to previous surveys of larger objects (Binzel et al. 2019, 2004; Perna et al. 2018), we report a lower fraction of S+Q-complex asteroids of 43.8 $\pm$ 4.6%. We associate this decrease with a lack of Phocaea family members at very small size. We also report higher fractions of X-complex and A-type asteroids of 23.8 $\pm$ 3.3% and 3.8 $\pm$ 1.3% respectively due to an increase of Hungaria family objects at small size. We find a strong correlation between the Q/S ratio and perihelion distance. We suggest this correlation is due to planetary close encounters with Venus playing a major role in turning asteroids from S to Q-type. This hypothesis is supported by a similar correlation between the Q/S ratio and Venus MOID.

astro-ph.EP↗

arXiv2019

Maximizing LSST Solar System Science: Approaches, Software Tools, and Infrastructure Needs

The Large Synoptic Survey Telescope (LSST) is expected to increase known small solar system object populations by an order of magnitude or more over the next decade, enabling a broad array of transformative solar system science investigations to be performed. In this white paper, we discuss software tools and infrastructure that we anticipate will be needed to conduct these investigations and outline possible approaches for implementing them. Feedback from the community or contributions to future updates of this work are welcome. Our aim is for this white paper to encourage further consideration of the software development needs of the LSST solar system science community, and also to be a call to action for working to meet those needs in advance of the expected start of the survey in late 2022.

astro-ph.EP↗

arXiv2019

Spitzer Albedos of Near-Earth Objects

Thermal infrared observations are the most effective way to measure asteroid diameter and albedo for a large number of near-Earth objects. Major surveys like NEOWISE, NEOSurvey, ExploreNEOs, and NEOLegacy find a small fraction of high albedo objects that do not have clear analogs in the current meteorite population. About 8% of Spitzer-observed near-Earth objects have nominal albedo solutions greater than 0.5. This may be a result of lightcurve variability leading to an incorrect estimate of diameter or inaccurate absolute visual magnitudes. For a sample of 23 high albedo NEOs we do not find that their shapes are significantly different from the McNeill et al. (2019) near-Earth object shape distribution. We performed a Monte Carlo analysis on 1505 near-Earth objects observed by Spitzer, sampling the visible and thermal fluxes of all targets to determine the likelihood of obtaining a high albedo erroneously. Implementing the McNeill shape distribution for near-Earth objects, we provide an upper-limit on the geometric albedo of 0.5+/-0.1 for the near-Earth population.

astro-ph.EP↗

arXiv2019

A common origin for dynamically associated near-Earth asteroid pairs

Though pairs of dynamically associated asteroids in the Main Belt have been identified and studied for over a decade, very few pair systems have been identified in the near-Earth asteroid population. We present data and analysis that supports the existence of two genetically related pairs in near-Earth space. The members of the individual systems, 2015 EE7 -- 2015 FP124 and 2017 SN16 -- 2018 RY7, are found to be of the same spectral taxonomic class, and both pairs are interpreted to have volatile-poor compositions. In conjunction with dynamical arguments, this suggests that these two systems formed via YORP spin-up and/or dissociation of a binary precursor. Backwards orbital integrations suggest a separation age of <10 kyr for the pair 2017 SN16 -- 2018 RY7, making these objects amongst the youngest multiple asteroid systems known to date. A unique separation age was not realized for 2015 EE7 -- 2015 FP124 due to large uncertainties associated with these objects' orbits. Determining the ages of such young pairs is of great value for testing models of space weathering and asteroid spin-state evolution. As the NEO catalog continues to grow with current and future discovery surveys, it is expected that more NEO pairs will be found, thus providing an ideal laboratory for studying time dependent evolutionary processes that are relevant to asteroids throughout the Solar System.

astro-ph.EP↗

arXiv2019

First Results from the rapid-response spectrophotometric characterization of Near-Earth Objects

As part of our multi-observatory, multi-filter campaign, we present \rmi color observations of 82 Near-Earth Objects (NEOs) obtained with the RATIR instrument on the 1.5m robotic telescope at the San Pedro Martir's National Observatory in Mexico. Our project is particularly focused on rapid response observations of small ($\lesssim 850$ m) NEOs. The rapid response and the use of spectrophotometry allows us to constrain the taxonomic classification of NEOs with high efficiency. Here we present the methodology of our observations and our result, suggesting that the ratio of C-type to S-type asteroids in a size range of $\sim$30-850m is 1.1, which is in accordance with our previous results. We also find that 10$\%$ of all NEOs in our sample are neither C- nor S-type asteroids

astro-ph.EP↗

arXiv2019

Constraining the Shape Distribution of Near Earth Objects from Partial Lightcurves

In the absence of dense photometry for a large population of Near Earth Objects (NEOs), the best method of obtaining a shape distribution comes from sparse photometry and partial lightcurves. We have used 867 partial lightcurves obtained by Spitzer to determine a shape distribution for sub-kilometre NEOs. From this data we find a best fit average elongation $\frac{b}{a}=0.72 \pm 0.08$. We compare this result with a shape distribution obtained from 1869 NEOs in the same size range observed by Pan-STARRS 1 and find the Spitzer-obtained elongation to be in excellent agreement with this PS1 value of $\frac{b}{a}=0.70 \pm 0.10$. These values are also in agreement with literature values for $1<D<10$ km objects in the main asteroid belt, however, there is a size discrepancy between the two datasets. Using a smaller sample of NEOs in the size range $1<D<5$ km from PS1 data, we obtain an average axis ratio $b/a = 0.70 \pm 0.12$. This is more elongated than the shape distribution for main belt objects in the same size regime, although the current uncertainties are sizeable and this should be verified using a larger data set. As future large surveys come online it will be possible to observe smaller main belt asteroids to allow for better comparisons of different sub-kilometre populations.

astro-ph.EP↗

arXiv2019

Enabling Deep All-Sky Searches of Outer Solar System Objects

A foundational goal of the Large Synoptic Survey Telescope (LSST) is to map the Solar System small body populations that provide key windows into understanding of its formation and evolution. This is especially true of the populations of the Outer Solar System -- objects at the orbit of Neptune $r > 30$AU and beyond. In this whitepaper, we propose a minimal change to the LSST cadence that can greatly enhance LSST's ability to discover faint distant Solar System objects across the entire wide-fast-deep (WFD) survey area. Specifically, we propose that the WFD cadence be constrained so as to deliver least one sequence of $\gtrsim 10$ visits per year taken in a $\sim 10$ day period in any combination of $g, r$, and $i$ bands. Combined with advanced shift-and-stack algorithms (Whidden et al. 2019) this modification would enable a nearly complete census of the outer Solar System to $\sim 25.5$ magnitude, yielding $4-8$x more KBO discoveries than with single-epoch baseline, and enabling rapid identification and follow-up of unusual distant Solar System objects in $\gtrsim 5$x greater volume of space. These increases would enhance the science cases discussed in Schwamb et al. (2018) whitepaper, including probing Neptune's past migration history as well as discovering hypothesized planet(s) beyond the orbit of Neptune (or at least placing significant constraints on their existence).

astro-ph.EP↗

arXiv2019

astroquery: An Astronomical Web-Querying Package in Python

astroquery is a collection of tools for requesting data from databases hosted on remote servers with interfaces exposed on the internet, including those with web pages but without formal application program interfaces (APIs). These tools are built on the Python requests package, which is used to make HTTP requests, and astropy, which provides most of the data parsing functionality. astroquery modules generally attempt to replicate the web page interface provided by a given service as closely as possible, making the transition from browser-based to command-line interaction easy. astroquery has received significant contributions from throughout the astronomical community, including several significant contributions from telescope archives. astroquery enables the creation of fully reproducible workflows from data acquisition through publication. This paper describes the philosophy, basic structure, and development model of the astroquery package. The complete documentation for astroquery can be found at http://astroquery.readthedocs.io/.

astro-ph.IM↗

arXiv2018

Deep Drilling Fields for Solar System Science

We propose an ecliptic Deep Drilling Field that will discover some 10,000~small and faint Kuiper Belt Objects (KBOs) --- primitive rocky/icy bodies that orbit at the outside of our Solar System and uniquely record the processes of planetary system formation and evolution. The primary goals are to measure the KBO size and shape distributions down to 25~km, a size that probes both the early and ongoing evolution of this population. These goals can be met with around 10~hours total of on-sky time (five separate fields that are observed for 2.1~hours each). Additional science will result from downstream observations that provide colors and orbit refinement, for a total time request of 40~hours over the ten year LSST main survey.

astro-ph.IM↗

arXiv2018

A Northern Ecliptic Survey for Solar System Science

Making an inventory of the Solar System is one of the four fundamental science requirements for the Large Synoptic Survey Telescope (LSST). The current baseline footprint for LSST's main Wide-Fast-Deep (WFD) Survey observes the sky below 0$^\circ$ declination, which includes only half of the ecliptic plane. Critically, key Solar System populations are asymmetrically distributed on the sky: they will be entirely missed, or only partially mapped, if only the WFD occurs. We propose a Northern Ecliptic Spur (NES) mini survey, observing the northern sky up to +10$^\circ$ ecliptic latitude, to maximize Solar System science with LSST. The mini survey comprises a total area of $\sim$5800 deg$^2$/604 fields, with 255 observations/field over the decade, split between g,r, and z bands. Our proposed survey will 1) obtain a census of main-belt comets; 2) probe Neptune's past migration history, by exploring the resonant structure of the Kuiper belt and the Neptune Trojan population; 3) explore the origin of Inner Oort cloud objects and place significant constraints on the existence of a hypothesized planet beyond Neptune; and 4) enable precise predictions of KBO stellar occultations. These high-ranked science goals of the Solar System Science Collaboration are only achievable with this proposed northern survey.

astro-ph.EP↗

arXiv2018

A near-Sun Solar System Twilight Survey with LSST

We propose a LSST Solar System near-Sun Survey, to be implemented during twilight hours, that extends the seasonal reach of LSST to its maximum as fresh sky is uncovered at about 50 square degrees per night (1500 sq. deg. per lunation) in the morning eastern sky, and surveyable sky is lost at the same rate to the western evening sky due to the Earth's synodic motion. By establishing near-horizon fence post picket lines to the far west and far east we address Solar System science use cases (including Near Earth Objects, Interior Earth Objects, Potentially Hazardous Asteroids, Earth Trojans, near-Sun asteroids, sun-grazing comets, and dormant comets) as well as provide the first look and last look that LSST will have at the transient and variable objects within each survey field. This proposed near-Sun Survey will also maximize the overlap with the field of regard of the proposed NEOCam spacecraft that will be stationed at the Earth's L1 Lagrange point and survey near quadrature with the Sun. This will allow LSST to incidently follow-up NEOCam targets and vice-versa (as well as targets from missions such as Euclid), and will roughly correspond to the Earth's L4 and L5 regions.

astro-ph.EP↗

arXiv2018

Main Belt Asteroid Shape Distribution from Gaia DR2

Gaia Data Release 2 includes observational data for 14,099 pre-selected asteroids. From the sparsely sampled G band photometry, we derive lower-limit lightcurve amplitudes for 11,665 main belt asteroids in order to provide constraints on the distribution of shapes in the asteroid main belt. Assuming a triaxial shape model for each asteroid, defined through the axial aspect ratios a > b and b=c, we find an average b/a=0.80+-0.04 for the ensemble, which is in agreement with previous results. By combining the Gaia data with asteroid properties from the literature, we investigate possible correlations of the aspect ratio with size, semi-major axis, geometric albedo, and intrinsic color. Based on our model simulations, we find that main belt asteroids greater than 50 km in diameter on average have higher b/a aspect ratios (are rounder) than smaller asteroids. We furthermore find significant differences in the shape distribution of main belt asteroids as a function of the other properties that do not affect the average aspect ratios. We conclude that a more detailed investigation of shape distribution correlations requires a larger data sample than is provided in Gaia Data Release 2.

astro-ph.EP↗

arXiv2018

Infrared Lightcurves of Near Earth Objects

We present lightcurves and derive periods and amplitudes for a subset of 38 near earth objects (NEOs) observed at 4.5 microns with the IRAC camera on the the Spitzer Space Telescope, many of them having no previously reported rotation periods. This subset was chosen from about 1800 IRAC NEO observations as having obvious periodicity and significant amplitude. For objects where the period observed did not sample the full rotational period, we derived lower limits to these parameters based on sinusoidal fits. Lightcurve durations ranged from 42 to 544 minutes, with derived periods from 16 to 400 minutes. We discuss the effects of lightcurve variations on the thermal modeling used to derive diameters and albedos from Spitzer photometry. We find that both diameters and albedos derived from the lightcurve maxima and minima agree with our previously published results, even for extreme objects, showing the conservative nature of the thermal model uncertainties. We also evaluate the NEO rotation rates, sizes, and their cohesive strengths.

astro-ph.EP↗