Multiwavelength Spectral Energy Distribution Research Articles

X-cigale: fitting AGN/galaxy SEDs from X-ray to infrared

ABSTRACT Code Investigating GALaxy Emission (cigale) is a powerful multiwavelength spectral energy distribution (SED) fitting code for extragalactic studies. However, the current version of cigale is not able to fit X-ray data, which often provide unique insights into active galactic nucleus (AGN) intrinsic power. We develop a new X-ray module for cigale, allowing it to fit SEDs from the X-ray to infrared (IR). We also improve the AGN fitting of cigale from UV-to-IR wavelengths. We implement a modern clumpy two-phase torus model, skirtor. To account for moderately extincted type 1 AGNs, we implement polar-dust extinction. We publicly release the source code (named ‘x-cigale’). We test x-cigale with X-ray detected AGNs in SDSS, COSMOS, and AKARI-NEP. The fitting quality (as indicated by reduced χ2) is good in general, indicating that x-cigale is capable of modelling the observed SED from X-ray to IR. We discuss constrainability and degeneracy of model parameters in the fitting of AKARI-NEP, for which excellent mid-IR photometric coverage is available. We also test fitting a sample of AKARI-NEP galaxies for which only X-ray upper limits are available from Chandra observations, and find that the upper limit can effectively constrain the AGN SED contribution for some systems. Finally, using x-cigale, we assess the ability of Athena to constrain the AGN activity in future extragalactic studies.

Spitzer Catalog of Herschel-selected Ultrared Dusty Star-forming Galaxies

Abstract The largest Herschel extragalactic surveys, H-ATLAS and HerMES, have selected a sample of “ultrared” dusty star-forming galaxies (DSFGs) with rising SPIRE flux densities (S 500 > S 350 > S 250; the so-called “500 μm risers”) as an efficient way for identifying DSFGs at higher redshift (z > 4). In this paper, we present a large Spitzer follow-up program of 300 Herschel ultrared DSFGs. We have obtained high-resolution Atacama Large Millimeter/submillimeter Array, Northern Extended Millimeter Array, and SMA data for 63 of them, which allow us to securely identify the Spitzer/IRAC counterparts and classify them as gravitationally lensed or unlensed. Within the 63 ultrared sources with high-resolution data, ∼65% appear to be unlensed and ∼27% are resolved into multiple components. We focus on analyzing the unlensed sample by directly performing multiwavelength spectral energy distribution modeling to derive their physical properties and compare with the more numerous z ∼ 2 DSFG population. The ultrared sample has a median redshift of 3.3, stellar mass of 3.7 × 1011 M ⊙, star formation rate (SFR) of 730 M ⊙ yr−1, total dust luminosity of 9.0 × 1012 L ⊙, dust mass of 2.8 × 109 M ⊙, and V-band extinction of 4.0, which are all higher than those of the ALESS DSFGs. Based on the space density, SFR density, and stellar mass density estimates, we conclude that our ultrared sample cannot account for the majority of the star-forming progenitors of the massive, quiescent galaxies found in infrared surveys. Our sample contains the rarer, intrinsically most dusty, luminous, and massive galaxies in the early universe that will help us understand the physical drivers of extreme star formation.

X-ray and optical observations of the black hole candidate MAXI J1828−249

AbstractWe report results from X-ray and optical observations of the Galactic black hole candidate MAXI J1828−249 performed with Suzaku and the Kanata telescope around the X-ray flux peak in the 2013 outburst. The time-averaged X-ray spectrum covering 0.6–168 keV was approximately characterized by a strong multi-color disk blackbody component with an inner disk temperature of ∼0.6 keV, and a power-law tail with a photon index of ∼2.0. We detected an additional structure at 5–10 keV, which can be modeled neither with X-ray reflection on the disk nor relativistic broadening of the disk emission. Instead, it was successfully reproduced with a Comptonization of disk photons by thermal electrons with a relatively low temperature (≲10 keV). We infer that the source was in the intermediate state, considering its long-term trend in the hardness intensity diagram, the strength of the spectral power-law tail, and its variability properties. The low-temperature Comptonization component could be produced in a boundary region between the truncated standard disk and the hot inner flow, or a Comptonizing region that somehow developed above the disk surface. The multi-wavelength spectral energy distribution suggests that the optical and ultraviolet fluxes were dominated by irradiated outer disk emission.

The unique case of the AGN core of M87: a misaligned low power blazar?

Abstract M87 hosts one of the closest jetted active galactic nucleus (AGN) to Earth. Thanks to its vicinity and to the large mass of is central black hole, M87 is the only source in which the jet can be directly imaged down to near-event horizon scales with radio very large baseline interferometry (VLBI). This property makes M87 a unique source to isolate and study jet launching, acceleration and collimation. In this paper we employ a multi-zone model designed as a parametrisation of general relativistic magneto-hydrodynamics (GRMHD); for the first time we reproduce the jet’s observed shape and multi-wavelength spectral energy distribution (SED) simultaneously. We find strong constraints on key physical parameters of the jet, such as the location of particle acceleration and the kinetic power. However, we under-predict the (unresolved) γ-ray flux of the source, implying that the high-energy emission does not originate in the magnetically-dominated inner jet regions. Our results have important implications both for comparisons of GRMHD simulations with observations, and for unified models of AGN classes.

Investigating the Stellar Mass Growth Histories of Satellite Galaxies as a Function of Infall Time Using Phase-space

Abstract We compile a large sample of nearby galaxies that are satellites of hosts using a well-known Sloan Digital Sky Survey (SDSS) group catalog. From this sample, we create an “ancient infallers” and “recent infallers” subsample, based on the mean infall time predicted from cosmological simulations for galaxies with their location in phase-space. We compare the stellar mass growth histories of the galaxies in these two subsamples, as determined from multiwavelength spectral energy distribution fitting that uses a comprehensive library of star formation history shapes derived from cosmological simulations. By simultaneously controlling for satellite stellar mass and host halo mass, we can clearly see the impact of time spent in their hosts. As we might predict, the ancient infaller population shows clear signs of earlier quenching, especially for lower mass satellites in more massive hosts. More importantly, we find that the effects are not limited to massive hosts. We find that hosts that might be considered low-mass groups (halo masses ∼1013 M ⊙) can significantly alter their satellites, even for massive satellites (stellar masses ∼1010 M ⊙). Intriguingly, we see changes in the mass growth history of the satellites of clusters as early as 8 or 9 Gyr ago, when they had not yet entered the virial radius of their current host. We propose that this could be the result of galaxies being preprocessed in low-mass substructures in the protocluster outskirts, prior to infall.

A multi-scale exploration of a massive young stellar object

Context.The rarity of young massive stars combined with the fact that they are often deeply embedded has limited the understanding of the formation of stars larger than 8M⊙. Ground based mid-infrared (IR) interferometry is one way of securing the spatial resolution required to probe the circumstellar environments of massive young stellar objects (MYSOs). Given that the spatial-frequency coverage of such observations is often incomplete, direct-imaging can be supplementary to such a dataset. By consolidating these observations with modelling, the features of a massive protostellar environment can be constrained.Aims.This paper aims to detail the physical characteristics of the protostellar environment of the MYSO G305.20+0.21 at three size-scales by fitting one 2.5D radiative transfer model to three different types of observations simultaneously, providing an extensive view of the accreting regions of the MYSO.Methods.Interferometry, imaging and a multi-wavelength spectral energy distribution (SED) are combined to study G305.20+0.21. The high-resolution observations were obtained using the Very Large Telescope’s MIDI and VISIR instruments, producing visibilities in theN-band and near-diffraction-limited imaging in theQ-band respectively. By fitting simulated observables, derived from the radiative transfer model, to our observations the properties of the MYSO are constrained.Results.The VISIR image shows elongation at 100 mas scales and also displays a degree of asymmetry. From the simulated observables derived from the radiative transfer model output we find that a central protostar with a luminosity of ~5 × 104L⊙surrounded by a low-density bipolar cavity, a flared 1M⊙disk and an envelope is sufficient to fit all three types of observational data for G305.20+0.21. The weak silicate absorption feature within the SED requires low-density envelope cavities to be successfully fit and is an atypical characteristic in comparison to previously studied MYSOs.Conclusions.The fact that the presence of a dusty disk provides the best fit to the MIDI visibilities implies that this MYSO is following a scaled-up version of the low-mass star formation process. The low density, low extinction environment implies the object is a more evolved MYSO and this combined with large inner radius of the disk suggests that it could be an example of a transitional disk around an MYSO.

Detection of White Dwarf companions to Blue Straggler Stars from UVIT observations of M67

AbstractWe investigate the old open cluster M67 using ultraviolet photometric data of Ultra-Violet Imaging Telescope in multi-filter far-UV bands. M67, well known for the presence of several blue straggler stars (BSS), has been put to detailed tests to understand their formation pathways. Currently, there are three accepted formation channels: mass transfer due to Roche-lobe overflow in binary systems, stellar mergers either due to dynamical collisions or through coalescence of close binaries. So far, there had not been any confirmed detection of a white dwarf (WD) companion to any of the BSSs in this cluster. Here, we present the detection of WD companions to 5 bright BSSs in M67. The multiwavelength spectral energy distributions covering 0.12 -11.5 μm range, were found to require binary spectral fits for 5 BSSs, consisting of a cool (BSS) and a hot companion. The parameters (Luminosity, Temperature, Radius and Mass) of the hot companions suggest them to be WDs with mass in the range 0.2 - 0.35 M⊙ with Teff ~11000 –24000 K.

Galaxy And Mass Assembly (GAMA): Timescales for galaxies crossing the green valley

We explore the constraints that can be placed on the evolutionary timescales for typical low redshift galaxies evolving from the blue cloud through the green valley and onto the red sequence. We utilise galaxies from the GAMA survey with 0.1 < z < 0.2 and classify them according to the intrinsic (u-r?) colours of their stellar populations, as determined by fits to their multi-wavelength spectral energy distributions. Using these fits to also determine stellar population ages and star formation timescales, we argue that our results are consistent with a green valley population dominated by galaxies that are simply decreasing their star formation (running out of gas) over a timescale of 2-4 Gyr which are seen at a specific epoch in their evolution (approximately 1.6 e-folding times after their peak in star formation). If their fitted star formation histories are extrapolated forwards, the green galaxies will further redden over time, until they attain the colours of a passive population. In this picture, no specific quenching event which cuts-off their star formation is required, though it remains possible that the decline in star formation in green galaxies may be expedited by internal or external forces. However, there is no evidence that green galaxies have recently changed their star formation timescales relative to their previous longer term star formation histories.

Breaking degeneracy in jet dynamics: multi-epoch joint modelling of the BL Lac PKS 2155–304

Supermassive black holes can launch powerful jets which can be some of the most luminous multi-wavelength sources; decades after their discovery their physics and energetics are still poorly understood. The past decade has seen a dramatic improvement in the quality of available data, but despite this improvement the semi-analytical modelling of jets has advanced slowly: simple one-zone models are still the most commonly employed method of interpreting data, in particular for AGN jets. These models can roughly constrain the properties of jets but they cannot unambiguously couple their emission to the launching regions and internal dynamics, which can be probed with simulations. However, simulations are not easily comparable to observations because they cannot yet self-consistently predict spectra. We present an advanced semi-analytical model which accounts for the dynamics of the whole jet, starting from a simplified parametrization of Relativistic Magnetohydrodynamics in which the magnetic flux is converted into bulk kinetic energy. To benchmark the model we fit six quasisimultaneous, multi-wavelength spectral energy distributions of the BL Lac PKS 2155-304 obtained by the TANAMI program, and we address the degeneracies inherent to such a complex model by employing a state-of-the-art exploration of parameter space, which so far has been mostly neglected in the study of AGN jets. We find that this new approach is much more effective than a single-epoch fit in providing meaningful constraints on model parameters.

The relationship between dust and [C i] atz = 1 and beyond

Measuring molecular gas mass is vital for understanding the evolution of galaxies at high redshifts (z$\geq$1). Most measurements rely on CO as a tracer, but dependences on metallicity, dynamics and surface density lead to systematic uncertainties in high-z galaxies, where these physical properties are difficult to observe, and where the physical environments can differ systematically from those at z=0. Dust continuum emission provides a potential alternative assuming a known dust/gas ratio, but this must be calibrated on a direct gas tracer at z$\geq$1. In this paper we consider the [CI] 492-GHz emission line, which has been shown to trace molecular gas closely throughout Galactic clouds and has the advantages of being optically thin in typical conditions (unlike CO), and being observable at accessible frequencies at high redshifts (in contrast to the low-excitation lines of CO). We use the Atacama Large Millimetre/submillimetre Array (ALMA) to measure [CI], CO(4-3) and dust emission in a representative sample of star-forming galaxies at z=1, and combine these data with multi-wavelength spectral energy distributions to study relationships between dust and gas components of galaxies. We uncover a strong [CI]-dust correlation, suggesting that both trace similar phases of the gas. By incorporating other samples from the literature, we show that this correlation persists over a wide range of luminosities and redshifts up to z$\sim$4. Finally we explore the implications of our results as an independent test of literature calibrations for dust as a tracer of gas mass, and for predicting the CI abundance.

A Luminous Transient Event in a Sample of WISE-selected Variable AGNs

Abstract Recently Assef et al. presented two catalogs of active galactic nucleus (AGN) candidates over 30,093 deg2 selected from the Wide-field Infrared Survey Explorer (WISE) observations. From their most reliable sample, Assef et al. identified 45 AGN candidates with the highest variability levels in the AllWISE catalog but that are not blazars. Here we present new spectroscopic observations of some of these targets to further constrain their nature. We also study their optical light curves using observations from the Catalina Real-Time Transient Survey (CRTS) and find that only seven show significant optical variability, and that five of those seven are spectroscopically classified as AGNs. In one of them, WISEA J094806.56+031801.7 (W0948+0318), we identify a transient event in the CRTS light curve. We present a detailed analysis of this transient and characterize it through its CRTS light curve and its multiwavelength spectral energy distribution obtained from GALEX, Pan-STARRS, and WISE observations. We find that the most likely source of the transient is a superluminous supernova (SLSN) in W0948+0318. We estimate the total radiated energy to be E = (1.6 ± 0.3) × 1052 erg, making it one of the most energetic SLSNe observed. Based on the lack of change in mid-IR color throughout and after the transient event, we speculate that the location of the SLSN is within the torus of the AGN. We identify nine possible analogs to W0948+0318 based on their WISE light curves. None show optically detected transients and hence suggest significant dust obscuration. Finally, we estimate a rate of &gt;2 × 10−7 yr−1 per AGN for these transients under the conservative assumption that none of the identified analogs have a common origin with the transient in W0948+0318.

Fermi-Large Area Telescope Observations of the Brightest Gamma-Ray Flare Ever Detected from CTA 102

Abstract We present a multiwavelength study of the FSRQ CTA 102 using Fermi-LAT and simultaneous Swift-XRT/UVOT observations. The Fermi-LAT telescope detected one of the brightest flares from this object during 2016 September to 2017 March. In the 190 days of the observation period, the source underwent four major flares. A detailed analysis of the temporal and spectral properties of these flares indicates the flare at MJD 57751.594 has a γ-ray flux of (30.12 ± 4.48) × 10−6 ph cm−2 s−1 (from 90 minutes binning) in the energy range of 0.1–300 GeV. This has been found to be the highest flux ever detected from CTA 102. Time dependent leptonic modeling of the preflare, rising state, flares and decaying state has been done. A single emission region of size 6.5 × 1016 cm has been used in our work to explain the multiwavelength spectral energy distributions. During flares, the luminosity in electrons increases nearly seventy times compared to the preflare state.

Detection of the blazar S4 0954+65 at very-high-energy with the MAGIC telescopes during an exceptionally high optical state

Aims. The very high energy (VHE ≳100 GeV) γ-ray MAGIC observations of the blazar S4 0954+65, were triggered by an exceptionally high flux state of emission in the optical. This blazar has a disputed redshift of z = 0.368 or z ≥ 0.45 and an uncertain classification among blazar subclasses. The exceptional source state described here makes for an excellent opportunity to understand physical processes in the jet of S4 0954+65 and thus contribute to its classification. Methods. We investigated the multiwavelength (MWL) light curve and spectral energy distribution (SED) of the S4 0954+65 blazar during an enhanced state in February 2015 and have put it in context with possible emission scenarios. We collected photometric data in radio, optical, X-ray, and γ-ray. We studied both the optical polarization and the inner parsec-scale jet behavior with 43 GHz data. Results. Observations with the MAGIC telescopes led to the first detection of S4 0954+65 at VHE. Simultaneous data with Fermi-LAT at high energy γ-ray (HE, 100 MeV &lt; E &lt; 100 GeV) also show a period of increased activity. Imaging at 43 GHz reveals the emergence of a new feature in the radio jet in coincidence with the VHE flare. Simultaneous monitoring of the optical polarization angle reveals a rotation of approximately 100°. Conclusions. The high emission state during the flare allows us to compile the simultaneous broadband SED and to characterize it in the scope of blazar jet emission models. The broadband spectrum can be modeled with an emission mechanism commonly invoked for flat spectrum radio quasars (FSRQs), that is, inverse Compton scattering on an external soft photon fieldfrom the dust torus, also known as external Compton. The light curve and SED phenomenology is consistent with an interpretation of a blob propagating through a helical structured magnetic field and eventually crossing a standing shock in the jet, a scenario typically applied to FSRQs and low-frequency peaked BL Lac objects (LBL).

SOFIA/HAWC+ Detection of a Gravitationally Lensed Starburst Galaxy at z = 1.03

Abstract We present the detection at 89 μm (observed frame) of the Herschel-selected gravitationally lensed starburst galaxy HATLAS J1429-0028 (also known as G15v2.19) in 15 minutes with the High-resolution Airborne Wideband Camera-plus (HAWC+) onboard the Stratospheric Observatory for Infrared Astronomy (SOFIA). The spectacular lensing system consists of an edge-on foreground disk galaxy at z = 0.22 and a nearly complete Einstein ring of an intrinsic ultra-luminous infrared (IR) galaxy at z = 1.03. Is this high IR luminosity powered by pure star formation (SF) or also an active galactic nucleus (AGN)? Previous nebular line diagnostics indicate that it is star formation dominated. We perform a 27-band multiwavelength spectral energy distribution (SED) modeling including the new SOFIA/HAWC+ data to constrain the fractional AGN contribution to the total IR luminosity. The AGN fraction in the IR turns out to be negligible. In addition, J1429-0028 serves as a testbed for comparing SED results from different models/templates and SED codes (magphys, sed3fit, and cigale). We stress that star formation history is the dominant source of uncertainty in the derived stellar mass (as high as a factor of ∼10) even in the case of extensive photometric coverage. Furthermore, the detection of a source at z ∼ 1 with SOFIA/HAWC+ demonstrates the potential of utilizing this facility for distant galaxy studies including the decomposition of SF/AGN components, which cannot be accomplished with other current facilities.

Protostellar classification using supervised machine learning algorithms

Classification of young stellar objects (YSOs) into different evolutionary stages helps us to understand the formation process of new stars and planetary systems. Such classification has traditionally been based on spectral energy distributions (SEDs). An alternative approach is provided by supervised machine learning algorithms. We attempt to classify a sample of Orion YSOs into different classes, where each source has already been classified using multiwavelength SED analysis. We used 8 different learning algorithms to classify the target YSOs, namely a decision tree, random forest, gradient boosting machine (GBM), logistic regression, na\"ive Bayes classifier, $k$-nearest neighbour classifier, support vector machine, and neural network. The classifiers were trained and tested by using a 10-fold cross-validation procedure. As the learning features, we employed ten continuum flux densities spanning from the near-IR to submm. With a classification accuracy of 82% (with respect to the SED-based classes), a GBM algorithm was found to exhibit the best performance. The lowest accuracy of 47% was obtained with a na\"ive Bayes classifier. Our analysis suggests that the inclusion of the 3.6 $\mu$m and 24 $\mu$m flux densities is useful to maximise the YSO classification accuracy. Although machine learning has the potential to provide a rapid and fairly reliable way to classify YSOs, an SED analysis is still needed to derive the physical properties of the sources, and to create the labelled training data. The classification accuracies can be improved with respect to the present results by using larger data sets, more detailed missing value imputation, and advanced ensemble methods. Overall, the application of machine learning is expected to be very useful in the era of big astronomical data, for example to quickly assemble interesting target source samples for follow-up studies.

Leptonic and Hadronic Modeling of Fermi-LAT Hard Spectrum Quasars and Predictions for High-energy Polarization

Abstract We present the results of a study of the time-averaged spectral energy distributions (SEDs) of eight flat spectrum radio quasars (FSRQs) present in the second catalog of high energy sources detected beyond 50 GeV by the Fermi Large Area Telescope (2FHL). Both leptonic and hadronic scenarios are adopted to explain the multiwavelength SEDs and we find them to be marginally consistent with the 2FHL spectra above 50 GeV. We derive the expected degree of X-ray and γ-ray polarizations both for the average and elevated activity states and note that (i) a hadronic radiative model consistently predicts a higher degree of high energy polarization compared to leptonic ones and (ii) the X-ray polarization degree is higher than the γ-ray polarization in the leptonic scenario, but similar to the γ-ray polarization if the observed radiation is powered by hadronic processes. From the leptonic modeling, the location of the γ-ray emitting region is found to be at the outer edge of the broad line region (BLR) and is consistent with the γγ opacity estimates for the γ-ray absorption by the BLR. We conclude that a majority of the FSRQs could be detected by the upcoming Cherenkov Telescope Array, though future high energy polarimeters will be able to detect them only during elevated activity states, which could provide supportive evidence for the hadronic origin of the X-ray and γ-ray emission.

Verification of the diffusive shock acceleration in Mrk 501

The present work considers a plane shock front propagating along a cylindrical jet. Electrons experience the diffusive shock acceleration around the shock front, and subsequently drift away into the downstream flow in which they emit most of their energy. Assuming a proper boundary condition at the interface between the shock zone and the downstream zone, we solve the transport equation for the electrons in the downstream flow zone, where the combined effects of escape, synchrotron and IC cooling in the Thomson regime are taken into account. Using the electron spectrum obtained in this manner we calculate the multi-wavelength spectral energy distribution of Mrk 501 in the synchrotron self-Compton scenario. We check numerically if the Klein-Nishina cross-section could be approximated to the Thomsom regime. We consider whether the model results yield physically reasonable parameters, and further discuss some of implications of the model results. It suggests that the process of diffusive shock acceleration operates in the outflow of Mrk 501.

Mr-Moose: an advanced SED-fitting tool for heterogeneous multi-wavelength data sets

We present the public release of MrMoose, a fitting procedure that is able to perform multi-wavelength and multi-object spectral energy distribution (SED) fitting in a Bayesian framework. This procedure is able to handle a large variety of cases, from an isolated source to blended multi-component sources from an heterogeneous dataset (i.e. a range of observation sensitivities and spectral/spatial resolutions). Furthermore, MrMoose handles upper-limits during the fitting process in a continuous way allowing models to be gradually less probable as upper limits are approached. The aim is to propose a simple-to-use, yet highly-versatile fitting tool fro handling increasing source complexity when combining multi-wavelength datasets with fully customisable filter/model databases. The complete control of the user is one advantage, which avoids the traditional problems related to the "black box" effect, where parameter or model tunings are impossible and can lead to overfitting and/or over-interpretation of the results. Also, while a basic knowledge of Python and statistics is required, the code aims to be sufficiently user-friendly for non-experts. We demonstrate the procedure on three cases: two artificially-generated datasets and a previous result from the literature. In particular, the most complex case (inspired by a real source, combining Herschel, ALMA and VLA data) in the context of extragalactic SED fitting, makes MrMoose a particularly-attractive SED fitting tool when dealing with partially blended sources, without the need for data deconvolution.

Detection of GeV Gamma-Ray Emission in the Direction of HESS J1731-347 with Fermi-LAT

Abstract We report the detection of GeV γ-ray emission from supernova remnant HESS J1731-347 using 9 yr of Fermi Large Area Telescope data. We find a slightly extended GeV source in the direction of HESS J1731-347. The spectrum above 1 GeV can be fitted by a power law with an index of Γ = 1.77 ± 0.14, and the GeV spectrum connects smoothly with the TeV spectrum of HESS J1731-347. Either a hadronic–leptonic or a pure leptonic model can fit the multiwavelength spectral energy distribution of the source. However, the hard GeV γ-ray spectrum is more naturally produced in a leptonic (inverse Compton scattering) scenario, under the framework of diffusive shock acceleration. We also searched for the GeV γ-ray emission from the nearby TeV source HESS J1729-345. No significant GeV γ-ray emission is found, and upper limits are derived.

The multiwavelength spectrum of NGC 3115: hot accretion flow properties

NGC 3115 is the nearest galaxy hosting a billion solar mass black hole and is also a low-luminosity active galactic nucleus (LLAGN). X-ray observations of this LLAGN are able to spatially resolve the hot gas within the sphere of gravitational influence of the supermassive black hole. These observations make NGC 3115 an important testbed for black hole accretion theory in galactic nuclei since they constrain the outer boundary conditions of the hot accretion flow. We present a compilation of the multiwavelength spectral energy distribution (SED) of the nucleus of NGC 3115 from radio to X-rays. We report the results from modeling the observed SED with radiatively inefficient accretion flow (RIAF) models. The radio emission can be well-explained by synchrotron emission from the RIAF without the need for contribution from a relativistic jet. We obtain a tight constraint on the RIAF density profile, $\rho(r) \propto r^{-0.73_{-0.02} ^{+0.01}}$, implying that mass-loss through subrelativistic outflows from the RIAF is significant. The lower frequency radio observation requires the synchrotron emission from a nonthermal electron population in the RIAF, similarly to Sgr A*.