-
On the importance of learning non-local dynamics for stable data-driven climate modeling: A 1D gravity wave-QBO testbed
Authors:
Hamid A. Pahlavan,
Pedram Hassanzadeh,
M. Joan Alexander
Abstract:
Machine learning (ML) techniques, especially neural networks (NNs), have shown promise in learning subgrid-scale (SGS) parameterizations for climate modeling. However, a major problem with data-driven parameterizations, particularly those learned with supervised algorithms, is instability when integrated with numerical solvers of large-scale processes. Current remedies are often ad-hoc and lack a…
▽ More
Machine learning (ML) techniques, especially neural networks (NNs), have shown promise in learning subgrid-scale (SGS) parameterizations for climate modeling. However, a major problem with data-driven parameterizations, particularly those learned with supervised algorithms, is instability when integrated with numerical solvers of large-scale processes. Current remedies are often ad-hoc and lack a theoretical foundation. Here, we combine ML theory and climate physics to address a source of instability in NN-based parameterization. We demonstrate the importance of learning spatially non-local dynamics using a 1D model of the quasi-biennial oscillation (QBO) with gravity wave (GW) parameterization as a testbed. While common offline metrics fail to identify shortcomings in learning non-local dynamics, we show that the receptive field (RF)-the region of the input an NN uses to predict an output-can identify instability a-priori. We find that NN-based parameterizations that seem to accurately predict GW forcings from wind profiles ($\mathbf{R^2 \approx 0.99}$) cause unstable simulations when RF is too small to capture the non-local dynamics, while NNs of the same size but large-enough RF are stable. Some architectures, e.g., Fourier neural operators, have inherently large RF. We also demonstrate that learning non-local dynamics can be crucial for the stability and accuracy of a data-driven spatiotemporal emulator of the entire zonal wind field. Given the ubiquity of non-local dynamics in the climate system, we expect the use of effective RF, which can be computed for any NN architecture, to be important for many applications. This work highlights the need to integrate ML theory with physics for designing/analyzing data-driven algorithms for weather/climate modeling.
△ Less
Submitted 6 July, 2024;
originally announced July 2024.
-
Data Imbalance, Uncertainty Quantification, and Generalization via Transfer Learning in Data-driven Parameterizations: Lessons from the Emulation of Gravity Wave Momentum Transport in WACCM
Authors:
Y. Qiang Sun,
Hamid A. Pahlavan,
Ashesh Chattopadhyay,
Pedram Hassanzadeh,
Sandro W. Lubis,
M. Joan Alexander,
Edwin Gerber,
Aditi Sheshadri,
Yifei Guan
Abstract:
Neural networks (NNs) are increasingly used for data-driven subgrid-scale parameterization in weather and climate models. While NNs are powerful tools for learning complex nonlinear relationships from data, there are several challenges in using them for parameterizations. Three of these challenges are 1) data imbalance related to learning rare (often large-amplitude) samples; 2) uncertainty quanti…
▽ More
Neural networks (NNs) are increasingly used for data-driven subgrid-scale parameterization in weather and climate models. While NNs are powerful tools for learning complex nonlinear relationships from data, there are several challenges in using them for parameterizations. Three of these challenges are 1) data imbalance related to learning rare (often large-amplitude) samples; 2) uncertainty quantification (UQ) of the predictions to provide an accuracy indicator; and 3) generalization to other climates, e.g., those with higher radiative forcing. Here, we examine the performance of methods for addressing these challenges using NN-based emulators of the Whole Atmosphere Community Climate Model (WACCM) physics-based gravity wave (GW) parameterizations as the test case. WACCM has complex, state-of-the-art parameterizations for orography-, convection- and frontal-driven GWs. Convection- and orography-driven GWs have significant data imbalance due to the absence of convection or orography in many grid points. We address data imbalance using resampling and/or weighted loss functions, enabling the successful emulation of parameterizations for all three sources. We demonstrate that three UQ methods (Bayesian NNs, variational auto-encoders, and dropouts) provide ensemble spreads that correspond to accuracy during testing, offering criteria on when a NN gives inaccurate predictions. Finally, we show that the accuracy of these NNs decreases for a warmer climate (4XCO2). However, the generalization accuracy is significantly improved by applying transfer learning, e.g., re-training only one layer using ~1% new data from the warmer climate. The findings of this study offer insights for develo** reliable and generalizable data-driven parameterizations for various processes, including (but not limited) to GWs.
△ Less
Submitted 27 November, 2023;
originally announced November 2023.
-
Characteristics of Gravity Waves in Opposing Phases of the QBO: A Reanalysis Perspective with ERA5
Authors:
Hamid A. Pahlavan,
John M. Wallace,
Qiang Fu,
M. Joan Alexander
Abstract:
ERA5 data for the period of 1979-2019 are used as a basis for investigating the properties of gravity waves as they disperse and propagate upward through the stratosphere during opposing phases of the QBO. Two-sided zonal wavenumber-frequency spectra of vertical velocity in the stratosphere exhibit distinctive gravity wave signatures. Consistent with theory, westward propagating waves tend to be s…
▽ More
ERA5 data for the period of 1979-2019 are used as a basis for investigating the properties of gravity waves as they disperse and propagate upward through the stratosphere during opposing phases of the QBO. Two-sided zonal wavenumber-frequency spectra of vertical velocity in the stratosphere exhibit distinctive gravity wave signatures. Consistent with theory, westward propagating waves tend to be suppressed during the easterly QBO phase and eastward propagating waves tend to be suppressed during the westerly phase. Cospectra of the vertical flux of zonal momentum also show significant asymmetries between eastward and westward propagating waves during opposing QBO phases. Phase speed spectra of the vertical flux of momentum are indicative of a strong dissipation of westward propagating gravity waves during the easterly phase and vice versa; i.e., a selective "wind filtering" of the waves as they propagate upward. The three-dimensional structure of the gravity waves is revealed by compositing. In the absence of a background zonal flow, the waves radiate outward and upward from their respective reference grid points in concentric rings. When a zonal flow is present, the rings are amplified and compressed upstream of the source and attenuated and stretched downstream of it, such that they assume the form of arcs. These results serve to confirm the applicability of the mechanism proposed by Lindzen and Holton (1968) to explain the downward propagation of the QBO. The QBO also influences the spectrum of gravity waves in ERA5 at 100 hPa, below the layer in which wind-filtering occurs.
△ Less
Submitted 17 September, 2023;
originally announced September 2023.
-
Explainable Offline-Online Training of Neural Networks for Parameterizations: A 1D Gravity Wave-QBO Testbed in the Small-data Regime
Authors:
Hamid A. Pahlavan,
Pedram Hassanzadeh,
M. Joan Alexander
Abstract:
There are different strategies for training neural networks (NNs) as subgrid-scale parameterizations. Here, we use a 1D model of the quasi-biennial oscillation (QBO) and gravity wave (GW) parameterizations as testbeds. A 12-layer convolutional NN that predicts GW forcings for given wind profiles, when trained offline in a big-data regime (100-years), produces realistic QBOs once coupled to the 1D…
▽ More
There are different strategies for training neural networks (NNs) as subgrid-scale parameterizations. Here, we use a 1D model of the quasi-biennial oscillation (QBO) and gravity wave (GW) parameterizations as testbeds. A 12-layer convolutional NN that predicts GW forcings for given wind profiles, when trained offline in a big-data regime (100-years), produces realistic QBOs once coupled to the 1D model. In contrast, offline training of this NN in a small-data regime (18-months) yields unrealistic QBOs. However, online re-training of just two layers of this NN using ensemble Kalman inversion and only time-averaged QBO statistics leads to parameterizations that yield realistic QBOs. Fourier analysis of these three NNs' kernels suggests why/how re-training works and reveals that these NNs primarily learn low-pass, high-pass, and a combination of band-pass filters, consistent with the importance of both local and non-local dynamics in GW propagation/dissipation. These findings/strategies apply to data-driven parameterizations of other climate processes generally.
△ Less
Submitted 16 September, 2023;
originally announced September 2023.
-
A Catalog of New Spectroscopically Confirmed Massive OB Stars in Carina
Authors:
Michael J. Alexander,
Richard J. Hanes,
Matthew S. Povich,
M. Virginia McSwain
Abstract:
The Carina star-forming region is one of the largest in the Galaxy, and its massive star population is still being unveiled. The large number of stars combined with high, and highly variable, interstellar extinction makes it inherently difficult to find OB stars in this type of young region. We present the results of a spectroscopic campaign to study the massive star population of the Carina Nebul…
▽ More
The Carina star-forming region is one of the largest in the Galaxy, and its massive star population is still being unveiled. The large number of stars combined with high, and highly variable, interstellar extinction makes it inherently difficult to find OB stars in this type of young region. We present the results of a spectroscopic campaign to study the massive star population of the Carina Nebula, with the primary goal to confirm or reject previously identified Carina OB star candidates. A total of 141 known O- and B-type stars and 94 candidates were observed, of which 73 candidates had a high enough signal-to-noise ratio to classify. We find 23 new OB stars within the Carina Nebula, a 32% confirmation rate. One of the new OB stars has blended spectra and is suspected to be a double-lined spectroscopic binary (SB2). We also reclassify the spectral types of the known OB stars and discover nine new SB2s among this population. Finally, we discuss the spatial distribution of these new OB stars relative to known structures in the Carina NSebula.
△ Less
Submitted 4 November, 2016;
originally announced November 2016.
-
Predicting Gaia's Parallax Distance to the Cygnus OB2 Association with Eclipsing Binaries
Authors:
Daniel C. Kiminki,
Henry A. Kobulnicky,
Carlos A. Vargas Álvarez,
Michael J. Alexander,
Michael J. Lundquist
Abstract:
The Cygnus OB2 Association is one of the nearest and largest collections of massive stars in the Galaxy. Situated at the heart of the "Cygnus X" complex of star-forming regions and molecular clouds, its distance has proven elusive owing to the ambiguous nature of kinematic distances along this $\ell\simeq80$ degree sightline and the heavy, patchy extinction. In an effort to refine the three-dimens…
▽ More
The Cygnus OB2 Association is one of the nearest and largest collections of massive stars in the Galaxy. Situated at the heart of the "Cygnus X" complex of star-forming regions and molecular clouds, its distance has proven elusive owing to the ambiguous nature of kinematic distances along this $\ell\simeq80$ degree sightline and the heavy, patchy extinction. In an effort to refine the three-dimensional geometry of key Cygnus~X constituents, we have measured distances to four eclipsing double-lined OB-type spectroscopic binaries that are probable members of Cyg~OB2. We find distances of $1.33\pm0.17$, $1.32\pm0.07$, $1.44\pm0.18$, and $1.32\pm0.13$ kpc toward MT91~372, MT91~696, CPR2002~A36, and Schulte~3 respectively. We adopt a weighted average distance of 1.33$\pm$0.06~kpc. This agrees well with spectrophotometric estimates for the Association as a whole and with parallax measurements of protostellar masers in the surrounding interstellar clouds, thereby linking the ongoing star formation in these clouds with Cyg~OB2. We also identify Schulte 3C (O9.5V), a 4" visual companion to the 4.75 day binary Schulte~3(A+B), as a previously unrecognized Association member.
△ Less
Submitted 12 August, 2015;
originally announced August 2015.
-
An updated gamma-ray analysis of the Be-BH binary HD~215227
Authors:
Michael J. Alexander,
M. Virginia McSwain
Abstract:
We report an updated analysis of the gamma-ray source AGL J2241+4454 that was detected as a brief two-day flare in 2010 by the AGILE satellite. The high-energy emission of AGL J2241+4454 has been attributed to the binary system HD 215227, which consists of a Be star being orbited by a black hole making it the first known Be-black hole binary system. We have analyzed the AGILE data and find a gamma…
▽ More
We report an updated analysis of the gamma-ray source AGL J2241+4454 that was detected as a brief two-day flare in 2010 by the AGILE satellite. The high-energy emission of AGL J2241+4454 has been attributed to the binary system HD 215227, which consists of a Be star being orbited by a black hole making it the first known Be-black hole binary system. We have analyzed the AGILE data and find a gamma-ray flux of $(1.8\pm0.7)\times10^{-6}$ ph cm$^{-2}$ s$^{-1}$, in agreement with the initial report. Additionally, we examined data from the Fermi LAT over several time intervals including the two day flare, the folded orbital phase, and the entire mission ($\sim$6-years). We do not detect AGL J2241+4454 over any of these time periods with Fermi and find upper limits of $1.1\times10^{-7}$ ph cm$^{-2}$ s$^{-1}$ and $5.2\times10^{-10}$ ph cm$^{-2}$ s$^{-1}$ for the flare and the full mission, respectively. We conclude that the HD 215227 Be-black hole binary is not a true gamma-ray binary as previous speculated. While analyzing the Fermi data of the AGL J2241+4454 region, we discovered a previously unknown gamma-ray source with average flux of $(13.56\pm0.02)\times10^{-8}$ ph cm$^{-2}$ s$^{-1}$ that is highly variable on monthly timescales. We associate this emission with the known quasar 87GB 215950.2+503417.
△ Less
Submitted 25 February, 2015;
originally announced February 2015.
-
An All-Sky Sample of Intermediate-Mass Star-Forming Regions
Authors:
Michael J. Lundquist,
Henry A. Kobulnicky,
Michael J. Alexander,
Charles R. Kerton,
Kim Arvidsson
Abstract:
We present an all-sky sample of 984 candidate intermediate-mass Galactic star-forming regions color-selected from the Infrared Astronomical Satellite (IRAS) Point Source Catalog and morphologically classify each object using mid-infrared Wide-field Infrared Survey Explorer (WISE) images. Of the 984 candidates, 616 are probable star-forming regions (62.6%), 128 are filamentary structures (13.0%), 3…
▽ More
We present an all-sky sample of 984 candidate intermediate-mass Galactic star-forming regions color-selected from the Infrared Astronomical Satellite (IRAS) Point Source Catalog and morphologically classify each object using mid-infrared Wide-field Infrared Survey Explorer (WISE) images. Of the 984 candidates, 616 are probable star-forming regions (62.6%), 128 are filamentary structures (13.0%), 39 are point-like objects of unknown nature (4.0%), and 201 are galaxies (20.4%). We conduct a study of four of these regions, IRAS 00259+5625, IRAS 00420+5530, IRAS 01080+5717, and IRAS 05380+2020, at Galactic latitudes |b| > 5 degrees using optical spectroscopy from the Wyoming Infrared Observatory along with near-infrared photometry from the Two-Micron All Sky Survey to investigate their stellar content. New optical spectra, color-magnitude diagrams, and color-color diagrams reveal their extinctions, spectrophotometric distances, and the presence of small stellar clusters containing 20-78 solar masses of stars. These low-mass diffuse star clusters contain 65-250 stars for a typical initial mass function, including one or more mid-B stars as their most massive constituents. Using infrared spectral energy distributions we identify young stellar objects near each region and assign probable masses and evolutionary stages to the protostars. The total infrared luminosity lies in the range 190 to 960 solar luminosities, consistent with the sum of the luminosities of the individually identified young stellar objects.
△ Less
Submitted 3 February, 2014;
originally announced February 2014.
-
The Interstellar Bubbles of G38.9-0.4 and the Impact of Stellar Feedback on Star Formation
Authors:
Michael J. Alexander,
Henry A. Kobulnicky,
Charles R. Kerton,
Kim Arvidsson
Abstract:
We present a study of the star formation (SF) region G38.9-0.4 using publicly available multiwavelength Galactic Plane surveys from ground- and space-based observatories. This region is composed of four bright mid-IR bubbles and numerous infrared dark clouds. Two bubbles, N 74 and N 75, each host a star cluster anchored by a single O9.5V star. We identified 162 young stellar objects (YSOs) and cla…
▽ More
We present a study of the star formation (SF) region G38.9-0.4 using publicly available multiwavelength Galactic Plane surveys from ground- and space-based observatories. This region is composed of four bright mid-IR bubbles and numerous infrared dark clouds. Two bubbles, N 74 and N 75, each host a star cluster anchored by a single O9.5V star. We identified 162 young stellar objects (YSOs) and classify 54 as stage I, 7 stage II, 6 stage III, and 32 ambiguous. We do not detect the classical signposts of triggered SF, i.e., star-forming pillars or YSOs embedded within bubble rims. We conclude that feedback-triggered SF has not occurred in G38.9-0.4. The YSOs are preferentially coincident with infrared dark clouds. This leads to a strong correlation between areal YSO mass surface density and gas mass surface density with a power law slope near 1.3, which closely matches the Schmidt-Kennicutt Law. The correlation is similar inside and outside the bubbles and may mean that the SF efficiency is neither enhanced nor supressed in regions potentially influenced by stellar feedback. This suggests that gas density, regardless of how it is collected, is a more important driver of SF than stellar feedback. Larger studies should be able to quantify the fraction of all SF that is feedback-triggered by determining the fraction SF, feedback-compressed gas surrounding HII regions relative to that already present in molecular clouds.
△ Less
Submitted 26 April, 2013;
originally announced April 2013.
-
Massive Star Formation at the Periphery of the Evolved Giant HII Region W 39
Authors:
C. R. Kerton,
K. Arvidsson,
M. J. Alexander
Abstract:
We present the first detailed study of the large, ~30 pc diameter, inner-Galaxy HII region W 39. Radio recombination line observations combined with HI absorption spectra and Galactic rotation models show that the region lies at V(LSR) = +65.4+/-0.5 km/s corresponding to a near kinematic distance of 4.5+/-0.2 kpc. Analysis of radio continuum emission shows that the HII region is being powered by a…
▽ More
We present the first detailed study of the large, ~30 pc diameter, inner-Galaxy HII region W 39. Radio recombination line observations combined with HI absorption spectra and Galactic rotation models show that the region lies at V(LSR) = +65.4+/-0.5 km/s corresponding to a near kinematic distance of 4.5+/-0.2 kpc. Analysis of radio continuum emission shows that the HII region is being powered by a cluster of OB stars with a combined hydrogen-ionizing luminosity of log(Q) >=50, and that there are three compact HII regions located on the periphery of W 39, each with log(Q)~48.5 (single O7 - O9 V star equivalent). In the infrared, W 39 has a hierarchical bubble morphology, and is a likely site of sequential star formation involving massive stars. Kinematic models of the expansion of W 39 yield timescales of order Myr consistent with a scenario where the formation of the smaller HII regions has been triggered by the expansion of W 39. Using Spitzer GLIMPSE and MIPSGAL data we show that star-formation activity is not distributed uniformly around the periphery of W 39 but is concentrated in two areas that include the compact HII regions as well as a number of intermediate-mass Class I and Class II YSOs.
△ Less
Submitted 14 January, 2013;
originally announced January 2013.
-
Ultra-compact Embedded Clusters in the Galactic Plane
Authors:
Michael J. Alexander,
Henry A. Kobulnicky
Abstract:
We have identified a previously unrecognized population of very compact, embedded low-mass Galactic stellar clusters. These tight (r$ \approx $0.14 pc) grou**s appear as bright singular objects at the few arcsec resolution of the Spitzer Space Telescope at 8 and 24 $μ$m but become resolved in the sub-arcsecond UKIDSS images. They average six stars per cluster surrounded by diffuse infrared emiss…
▽ More
We have identified a previously unrecognized population of very compact, embedded low-mass Galactic stellar clusters. These tight (r$ \approx $0.14 pc) grou**s appear as bright singular objects at the few arcsec resolution of the Spitzer Space Telescope at 8 and 24 $μ$m but become resolved in the sub-arcsecond UKIDSS images. They average six stars per cluster surrounded by diffuse infrared emission and coincide with 100 -- 300 M$_{\sun}$ clumps of molecular material within a larger molecular cloud. The magnitudes of the brightest stars are consistent with mid- to early-B stars anchoring $\sim$80 M$_{\sun}$ star clusters. Their evolutionary descendants are likely to be Herbig Ae/Be pre-main sequence clusters. These ultra-compact embedded clusters (UCECs) may fill part of the low-mass void in the embedded cluster mass function. We provide an initial catalog of 18 UCECs drawn from infrared Galactic Plane surveys.
△ Less
Submitted 25 July, 2012;
originally announced July 2012.
-
A Fresh Catch of Massive Binaries in the Cygnus OB2 Association
Authors:
Henry A. Kobulnicky,
Rachel A. Smullen,
Daniel C. Kiminki,
Jessie C. Runnoe,
Earl S. Wood,
Garrett Long,
Michael J. Alexander,
Michael J. Lundquist,
Carlos A. Vargas-Alvarez
Abstract:
Massive binary stars may constitute a substantial fraction of progenitors to supernovae and gamma-ray bursts, and the distribution of their orbital characteristics holds clues to the formation process of massive stars. As a contribution to securing statistics on OB-type binaries, we report the discovery and orbital parameters for five new systems as part of the Cygnus OB2 Radial Velocity Survey. F…
▽ More
Massive binary stars may constitute a substantial fraction of progenitors to supernovae and gamma-ray bursts, and the distribution of their orbital characteristics holds clues to the formation process of massive stars. As a contribution to securing statistics on OB-type binaries, we report the discovery and orbital parameters for five new systems as part of the Cygnus OB2 Radial Velocity Survey. Four of the new systems (MT070, MT174, MT267, and MT734 (a.k.a. VI Cygni #11) are single-lined spectroscopic binaries while one (MT103) is a double-lined system (B1V+B2V). MT070 is noteworthy as the longest period system yet measured in Cyg OB2, with P=6.2 yr. The other four systems have periods ranging between 4 and 73 days. MT174 is noteworthy for having a probable mass ratio q<0.1, making it a candidate progenitor to a low-mass X-ray binary. These measurements bring the total number of massive binaries in Cyg OB2 to 25, the most currently known in any single cluster or association.
△ Less
Submitted 28 June, 2012;
originally announced June 2012.
-
A Sample of Intermediate-Mass Star-Forming Regions: Making Stars at Mass Column Densities <1 g/cm^2
Authors:
Kim Arvidsson,
Charles R. Kerton,
Michael J. Alexander,
Henry A. Kobulnicky,
Brian Uzpen
Abstract:
In an effort to understand the factors that govern the transition from low- to high-mass star formation, we identify for the first time a sample of intermediate-mass star-forming regions (IM SFRs) where stars up to - but not exceeding - 8 solar masses are being produced. We use IRAS colors and Spitzer Space Telescope mid-IR images, in conjunction with millimeter continuum and CO maps, to compile a…
▽ More
In an effort to understand the factors that govern the transition from low- to high-mass star formation, we identify for the first time a sample of intermediate-mass star-forming regions (IM SFRs) where stars up to - but not exceeding - 8 solar masses are being produced. We use IRAS colors and Spitzer Space Telescope mid-IR images, in conjunction with millimeter continuum and CO maps, to compile a sample of 50 IM SFRs in the inner Galaxy. These are likely to be precursors to Herbig AeBe stars and their associated clusters of low-mass stars. IM SFRs constitute embedded clusters at an early evolutionary stage akin to compact HII regions, but they lack the massive ionizing central star(s). The photodissociation regions that demarcate IM SFRs have typical diameters of ~1 pc and luminosities of ~10^4 solar luminosities, making them an order of magnitude less luminous than (ultra)compact HII regions. IM SFRs coincide with molecular clumps of mass ~10^3 solar masses which, in turn, lie within larger molecular clouds spanning the lower end of the giant molecular cloud mass range, 10^4-10^5 solar masses. The IR luminosity and associated molecular mass of IM SFRs are correlated, consistent with the known luminosity-mass relationship of compact HII regions. Peak mass column densities within IM SFRs are ~0.1-0.5 g/cm^2, a factor of several lower than ultra-compact HII regions, supporting the proposition that there is a threshold for massive star formation at ~1 g/cm^2.
△ Less
Submitted 18 June, 2010; v1 submitted 25 May, 2010;
originally announced May 2010.
-
The Discovery of a Massive Cluster of Red Supergiants with GLIMPSE
Authors:
Michael J. Alexander,
Henry A. Kobulnicky,
Dan P. Clemens,
Katherine Jameson,
April Pinnick,
Michael Pavel
Abstract:
We report the discovery of a previously unknown massive Galactic star cluster at l=29.22, b=-0.20. Identified visually in mid-IR images from the Spitzer GLIMPSE survey, the cluster contains at least 8 late-type supergiants, based on followup near-IR spectroscopy, and an additional 3-6 candidate supergiant embers having IR photometry consistent with a similar distance and reddening. The cluster l…
▽ More
We report the discovery of a previously unknown massive Galactic star cluster at l=29.22, b=-0.20. Identified visually in mid-IR images from the Spitzer GLIMPSE survey, the cluster contains at least 8 late-type supergiants, based on followup near-IR spectroscopy, and an additional 3-6 candidate supergiant embers having IR photometry consistent with a similar distance and reddening. The cluster lies at a local minimum in the 13-CO column density and 8 micron emission. We interpret this feature as a hole carved by the energetic winds of the evolving massive stars. The 13-CO hole seen in molecular maps at V_LSR ~95 km/s corresponds to near/far kinematic distances of 6.1/8.7+/-1 kpc. We calculate a mean spectrophotometric distance of 7.0^+3.7_-2.4 kpc, broadly consistent with the kinematic distances inferred. This location places it near the northern end of the Galactic bar. For the mean extinction of A_V=12.6+/-0.5 mag (A_K=1.5+/-0.1 mag), the color-magnitude diagram of probable cluster members is well fit by isochrones in the age range 18-24 Myr. The estimated cluster mass is ~20,000 Msun. With the most massive original cluster stars likely deceased, no strong radio emission is detected in this vicinity. As such, this RSG cluster is representative of adolescent massive Galactic clusters that lie hidden behind many magnitudes of dust obscuration. This cluster joins two similar red supergiant clusters as residents of the volatile region where the end of our Galaxy's bar joins the base of the Scutum-Crux piral arm, suggesting a recent episode of widespread massive star formation there.
△ Less
Submitted 13 March, 2009;
originally announced March 2009.
