Machine learning (ML) is now used in many areas of astrophysics, from detecting exoplanets in Kepler transit signals to removing telescope systematics. Recent work demonstrated the potential of using ...ML algorithms for atmospheric retrieval by implementing a random forest (RF) to perform retrievals in seconds that are consistent with the traditional, computationally expensive nested-sampling retrieval method. We expand upon their approach by presenting a new ML model, plan-net, based on an ensemble of Bayesian neural networks (BNNs) that yields more accurate inferences than the RF for the same data set of synthetic transmission spectra. We demonstrate that an ensemble provides greater accuracy and more robust uncertainties than a single model. In addition to being the first to use BNNs for atmospheric retrieval, we also introduce a new loss function for BNNs that learns correlations between the model outputs. Importantly, we show that designing ML models to explicitly incorporate domain-specific knowledge both improves performance and provides additional insight by inferring the covariance of the retrieved atmospheric parameters. We apply plan-net to the Hubble Space Telescope Wide Field Camera 3 transmission spectrum for WASP-12b and retrieve an isothermal temperature and water abundance consistent with the literature. We highlight that our method is flexible and can be expanded to higher-resolution spectra and a larger number of atmospheric parameters.
•Early Mars could have been warmed by a thick CO2–H2 greenhouse atmosphere.•Photochemical modeling is used to study conversion of reduced volcanic gases to H2.•Getting the minimum H2 requires high ...outgassing or additional hydrogen sources.•Oxidation of iron by serpentinization can help boost atmospheric H2 concentration.•The Curiosity Rover could test the H2–CO2 greenhouse hypothesis in a variety of ways.
A recent study by Ramirez et al. (Ramirez, R.M. et al. 2014. Nat. Geosci. 7(1), 59–63. <http://www.nature.com/doifinder/10.1038/ngeo2000> (accessed 16.09.14)) demonstrated that an atmosphere with 1.3–4bar of CO2 and H2O, in addition to 5–20% H2, could have raised the mean annual and global surface temperature of early Mars above the freezing point of water. Such warm temperatures appear necessary to generate the rainfall (or snowfall) amounts required to carve the ancient martian valleys. Here, we use our best estimates for early martian outgassing rates, along with a 1-D photochemical model, to assess the conversion efficiency of CO, CH4, and H2S to CO2, SO2, and H2. Our outgassing estimates assume that Mars was actively recycling volatiles between its crust and interior, as Earth does today. H2 production from serpentinization and deposition of banded iron-formations is also considered. Under these assumptions, maintaining an H2 concentration of ∼1–2% by volume is achievable, but reaching 5% H2 requires additional H2 sources or a slowing of the hydrogen escape rate below the diffusion limit. If the early martian atmosphere was indeed H2-rich, we might be able to see evidence of this in the rock record. The hypothesis proposed here is consistent with new data from the Curiosity Rover, which show evidence for a long-lived lake in Gale Crater near Mt. Sharp. It is also consistent with measured oxygen fugacities of martian meteorites, which show evidence for progressive mantle oxidation over time.
Over the last few years, a number of authors have suggested that, under certain circumstances, molecular oxygen (O2) or ozone (O3) generated by abiotic processes may accumulate to detectable ...concentrations in a habitable terrestrial planet's atmosphere, producing so-called "false positives" for life. But the models have occasionally disagreed with each other, with some predicting false positives, and some not, for the same apparent set of circumstances. We show here that photochemical false positives derive either from inconsistencies in the treatment of atmospheric and global redox balance or from the treatment (or lack thereof) of lightning. For habitable terrestrial planets with even trace amounts of atmospheric N2, NO produced by lightning catalyzes the recombination of CO and O derived from CO2 photolysis and should be sufficient to eliminate all reported false positives. Molecular oxygen thus remains a useful biosignature gas for Earth-like extrasolar planets, provided that the planet resides within the conventional liquid water habitable zone and has not experienced distinctly non-Earth-like, irrecoverable water loss.
Geological and biological evidence suggests that Earth was warm during most of its early history, despite the fainter young Sun. Upper bounds on the atmospheric CO2 concentration in the Late ...Archean/Paleoproterozoic (2.8-2.2 Ga) from paleosol data suggest that additional greenhouse gases must have been present. Methanogenic bacteria, which were arguably extant at that time, may have contributed to a high concentration of atmospheric CH4, and previous calculations had indicated that a CH4-CO2-H2O greenhouse could have produced warm Late Archean surface temperatures while still satisfying the paleosol constraints on pCO2. Here, we revisit this conclusion. Correction of an error in the CH4 absorption coefficients, combined with the predicted early onset of climatically cooling organic haze, suggest that the amount of greenhouse warming by CH4 was more limited and that pCO2 must therefore have been 0.03 bar, at or above the upper bound of the value obtained from paleosols. Enough warming from CH4 remained in the Archean, however, to explain why Earth's climate cooled and became glacial when atmospheric O2 levels rose in the Paleoproterozoic. Our new model also shows that greenhouse warming by higher hydrocarbon gases, especially ethane (C2H6), may have helped to keep the Late Archean Earth warm.
Understanding what processes govern atmospheric escape and the loss of planetary water is of paramount importance for understanding how life in the universe can exist. One mechanism thought to be ...important at all planets is an ambipolar electric field that helps ions overcome gravity. We report the discovery and first quantitative extraterrestrial measurements of such a field at the planet Venus. Unexpectedly, despite comparable gravity, we show the field to be five times stronger than in Earths similar ionosphere. Contrary to our understanding, Venus would still lose heavy ions (including oxygen and all water-group species) to space, even if there were no stripping by the solar wind. We therefore find that it is possible for planets to lose heavy ions to space entirely through electric forces in their ionospheres and such an electric wind must be considered when studying the evolution and potential habitability of any planet in any star system.
O
and O
have been long considered the most robust individual biosignature gases in a planetary atmosphere, yet multiple mechanisms that may produce them in the absence of life have been described. ...However, these abiotic planetary mechanisms modify the environment in potentially identifiable ways. Here we briefly discuss two of the most detectable spectral discriminants for abiotic O
/O
: CO and O
. We produce the first explicit self-consistent simulations of these spectral discriminants as they may be seen by
(
). If
-NIRISS and/or NIRSpec observe CO (2.35, 4.6
m) in conjunction with CO
(1.6, 2.0, 4.3
m) in the transmission spectrum of a terrestrial planet it could indicate robust CO
photolysis and suggest that a future detection of O
or O
might not be biogenic. Strong O
bands seen in transmission at 1.06 and 1.27
m could be diagnostic of a post-runaway O
-dominated atmosphere from massive H-escape. We find that for these false positive scenarios, CO at 2.35
m, CO
at 2.0 and 4.3
m, and O
at 1.27
m are all stronger features in transmission than O
/O
and could be detected with S/Ns ≳ 3 for an Earth-size planet orbiting a nearby M dwarf star with as few as 10 transits, assuming photon-limited noise. O
bands could also be sought in UV/VIS/NIR reflected light (at 0.345, 0.36, 0.38, 0.445, 0.475, 0.53, 0.57, 0.63, 1.06, and 1.27
m) by a next generation direct-imaging telescope such as LUVOIR/HDST or HabEx and would indicate an oxygen atmosphere too massive to be biologically produced.
VPLanet: The Virtual Planet Simulator Barnes, Rory; Luger, Rodrigo; Deitrick, Russell ...
Publications of the Astronomical Society of the Pacific,
02/2020, Letnik:
132, Številka:
1008
Journal Article
Recenzirano
Odprti dostop
We describe a software package called VPLanet that simulates fundamental aspects of planetary system evolution over Gyr timescales, with a focus on investigating habitable worlds. In this initial ...release, eleven physics modules are included that model internal, atmospheric, rotational, orbital, stellar, and galactic processes. Many of these modules can be coupled to simultaneously simulate the evolution of terrestrial planets, gaseous planets, and stars. The code is validated by reproducing a selection of observations and past results. VPLanet is written in C and designed so that the user can choose the physics modules to apply to an individual object at runtime without recompiling, i.e., a single executable can simulate the diverse phenomena that are relevant to a wide range of planetary and stellar systems. This feature is enabled by matrices and vectors of function pointers that are dynamically allocated and populated based on user input. The speed and modularity of VPLanet enables large parameter sweeps and the versatility to add/remove physical phenomena to assess their importance. VPLanet is publicly available from a repository that contains extensive documentation, numerous examples, Python scripts for plotting and data management, and infrastructure for community input and future development.
We use sulfur (S) isotope signatures within ancient sediments and a photochemical model of sulfur dioxide (SO
2) photolysis to interpret the evolution of the atmosphere over the first half of Earth's ...history. A decrease in mass-independent sulfur isotope fractionation has been reported in Archean rocks deposited between ~ 2.7 Ga and ~ 3.2 Ga, and is reinforced by new S isotope data that we report here. This pattern has been interpreted by some as evidence that atmospheric oxygen (O
2) was elevated during this time. In this paper, we argue against that conclusion, and show that it is inconsistent with other geochemical data. In its place, we propose a new model that can explain the sulfur isotope record that can also avoid conflicts with independent constraints on O
2 and account for concurrent glacial deposits. Specifically, we suggest that prior to the rise of O
2 excursions in the sulfur isotope record were modulated by the thickness of an organic haze. This haze would have blocked the lower atmosphere from the UV fluxes responsible for the anomalous sulfur photochemistry and would have caused an anti-greenhouse effect that triggered the glaciations. We used a photochemical model to verify that a haze could have affected the isotopic signal in this manner, and to examine how changes in atmospheric methane (CH
4) and carbon dioxide (CO
2) concentrations could have controlled haze thickness. Finally, we combined the resulting relationships with climate models and sulfur isotope and glacial records to deduce a new evolutionary sequence for Archean climate, surface chemistry, and biology.
Transit spectroscopy of terrestrial planets around nearby M dwarfs will be a primary goal of space missions in coming decades. Three-dimensional climate modeling has shown that slow-synchronous ...rotating terrestrial planets may develop thick clouds at the substellar point, increasing the albedo. For M dwarfs with Teff > 3000 K, such planets at the inner habitable zone (IHZ) have been shown to retain moist greenhouse conditions, with enhanced stratospheric water vapor (fH2O > 10−3) and low Earth-like surface temperatures. However, M dwarfs also possess strong UV activity, which may effectively photolyze stratospheric H2O. Prior modeling efforts have not included the impact of high stellar UV activity on the H2O. Here, we employ a 1D photochemical model with varied stellar UV, to assess whether H2O destruction driven by high stellar UV would affect its detectability in transmission spectroscopy. Temperature and water vapor profiles are taken from published 3D climate model simulations for an IHZ Earth-sized planet around a 3300 K M dwarf with an N2-H2O atmosphere; they serve as self-consistent input profiles for the 1D model. We explore additional chemical complexity within the 1D model by introducing other species into the atmosphere. We find that as long as the atmosphere is well-mixed up to 1 mbar, UV activity appears to not impact detectability of H2O in the transmission spectrum. The strongest H2O features occur in the James Webb Space Telescope MIRI instrument wavelength range and are comparable to the estimated systematic noise floor of ∼50 ppm.
PDF
