About me

I have moved! I am now a Flatiron Research Fellow at the Center for Computational Astrophysics in the Flatiron Institute in New York City. You can find my new personal webpage here.

Launching galactic winds

Galactic winds play a critically important role in galaxy evolution—they help limit galactic star formation efficiencies by expelling material from the interstellar medium and by halting gas inflow into galaxies, and they enrich and heat the circumgalactic medium. We use idealized global and local simulations to study how SNe drive winds.

Circumgalactic medium

Galaxies are surrounded by diffuse gas that fills their dark matter halos. The structure and kinematics of the CGM depend sensitively on halo mass (and possibly feedback properties). In recent years CGM observations have presented rich but surprising constraints that we use controlled simulations to try to shed light on.

Supernova driven galactic winds

Clustered Supernovae Drive Powerful Galactic Winds After Super-Bubble Breakout

Drummond Fielding, Eliot Quataert, and Davide Martizzi, submitted to MNRAS (2018).

We use three-dimensional hydrodynamic simulations of vertically stratified patches of galactic discs to study how the spatio-temporal clustering of supernovae (SNe) enhances the power of galactic winds. SNe that are randomly distributed throughout a galactic disc drive inefficient galactic winds because most supernova remnants lose their energy radiatively before breaking out of the disc. Accounting for the fact that most star formation is clustered alleviates this problem. Super-bubbles driven by the combined effects of clustered SNe propagate rapidly enough to break out of galactic discs well before the clusters’ SNe stop going off. The radiative losses post- breakout are reduced dramatically and a large fraction (≳ 0.2) of the energy released by SNe vents into the halo powering a strong galactic wind. These energetic winds are capable of providing strong preventative feedback and eject substantial mass from the galaxy with outflow rates on the order of the star formation rate. The momentum flux in the wind is only of order that injected by the SNe, because the hot gas vents before doing significant work on the surroundings. We show that our conclusions hold for a range of galaxy properties, both in the local Universe (e.g., M82) and at high redshift (e.g., z ∼ 2 star forming galaxies). We further show that if the efficiency of forming star clusters increases with increasing gas surface density, as suggested by theoretical arguments, the condition for star cluster-driven super-bubbles to break out of galactic discs corresponds to a threshold star formation rate surface density for the onset of galactic winds ∼ 0.03 M yr−1 kpc−2, of order that observed.

Many additional movies that show different quantities and simulations of different gas surface densities, cluster masses, and simulation setups can be found here.

How Supernovae Launch Galactic Winds

Drummond Fielding, Eliot Quataert, Davide Martizzi, and Claude-André Faucher-Giguère MNRAS, 470L, 39 (2017).

We use idealized three-dimensional hydrodynamic simulations of global galactic discs to study the launching of galactic winds by supernovae (SNe). The simulations resolve the cooling radii of the majority of supernova remnants (SNRs) and thus self-consistently capture how SNe drive galactic winds. We find that SNe launch highly supersonic winds with properties that agree reasonably well with expectations from analytic models. The energy loading (ηE = Ėwind / ĖSN) of the winds in our simulations are well converged with spatial resolution while the wind mass loading (ηM = Ṁwind / Ṁ∗) decreases with resolution at the resolutions we achieve. We present a simple analytic model based on the concept that SNRs with cooling radii greater than the local scale height breakout of the disc and power the wind. This model successfully explains the dependence (or lack thereof) of ηE (and by extension ηM) on the gas surface density, star formation efficiency, disc radius, and the clustering of SNe. The winds in the majority of our simulations are weaker than expected in reality, likely due to the fact that we seed SNe preferentially at density peaks. Clustering SNe in time and space substantially increases the wind power.

Supernova Feedback in a Vertically Stratified Medium: Interstellar Turbulence and Galactic Wind Launching

Davide Martizzi, Drummond Fielding, Claude-André Faucher-Giguère, and Eliot Quataert, , MNRAS, 459, 2311 (2016).

We use local Cartesian simulations with a vertical gravitational potential to study how supernova (SN) feedback in stratified galactic discs drives turbulence and launches galactic winds. Our analysis includes three disc models with gas surface densities ranging from Milky Way-like galaxies to gas-rich ultra-luminous infrared galaxies (ULIRGs), and two different SN driving schemes (random and correlated with local gas density). In order to isolate the physics of SN feedback, we do not include additional feedback processes. We find that, in these local box calculations, SN feedback excites relatively low mass-weighted gas turbulent velocity dispersions ~3-7 km/s and low wind mass loading factors, less than 1, in all the cases we study. The low turbulent velocities and wind mass loading factors predicted by our local box calculations are significantly below those suggested by observations of gas-rich and rapidly star-forming galaxies; they are also in tension with global simulations of disc galaxies regulated by stellar feedback. Using a combination of numerical tests and analytic arguments, we argue that local Cartesian boxes cannot predict the properties of galactic winds because they do not capture the correct global geometry and gravitational potential of galaxies. The wind mass loading factors are in fact not well-defined in local simulations because they decline significantly with increasing box height. More physically realistic calculations (e.g., including a global galactic potential and disc rotation) will likely be needed to fully understand disc turbulence and galactic outflows, even for the idealized case of feedback by SNe alone.

Supernova driven galactic winds

The Circumgalactic Medium

--- Drummond Fielding, Eliot Quataert, Michael McCourt, and Todd Thompson The Impact of Star Formation Feedback on the Circumgalactic Medium. MNRAS, 466, 3810 (2017).

In the modern paradigm for the formation of structure in our universe, matter over-densities imprinted during the big bang collapse to form ‘halos’ of dark matter. These dark matter halos are initially filled with hot gas heated to the virial temperature by an accretion shock. Though the gravitational process of structure formation is self-similar, pioneering work in the 1970’s recognized that the cooling of halo gas ultimately differentiates galaxies from more massive systems such as galaxy groups or clusters. In galaxies with sufficiently low mass, halo gas can in principle cool, accrete, and play a role in star formation. In galaxy clusters, on the other hand, halo gas cannot cool and instead persists for the age of the universe, essentially independent of the galaxy formation taking place at smaller radii. The transition between these two outcomes for halo gas occurs around a total mass of ∼1012 M, relatively independent of redshift. Thus, a seemingly small distinction related to atomic physics ultimately drives all of the marked observational differences between galaxies and galaxy clusters. The basic physics of halo gas therefore plays a primary, but relatively unexplored, role in our history of the universe and the types of structures that form within it.

Until recently, we had only indirect measurements of halo gas in galaxies. As a result, current cosmological simulations have been tuned to reproduce the stellar properties of galaxies, which are much better constrained from observations. As new observational techniques begin to probe the halo gas in galaxies, however, it will become increasingly important to improve the prescriptions for gas physics in our models. Many of the new observations—particularly the HST COS quasar absorption spectra observations (e.g., Werk et al. 2014)—are at odds with the predictions from state-of-the-art cosmological simulations. Predicting, observing, and understanding the properties of halo gas represents the new frontier in galaxy formation.

In our first paper on this topic, we use idealized three-dimensional hydrodynamic simulations to study the dynamics and thermal structure of the circumgalactic medium (CGM) resulting from the interplay of cooling and stellar feedback in the presence of cosmological accretion of gas. As in the pioneering calculations without stellar feedback (e.g., Birnboim & Dekel 2003; Keres et al 2005), we find that above a critical halo mass (∼1012 M) halo gas is supported by thermal pressure created in the halo virial shock. The thermal properties of the halo gas at small radii (near any central galaxy) are regulated by feedback triggered when tcool/tff ~ 1−10 in the hot halo gas. Below a critical halo mass, ∼1012 M, however, there is no thermally supported halo. Instead, the halo gas properties are determined by the interaction between inflowing gas and galactic winds. The halo gas is not in thermal pressure balance, but is instead effectively supported by random motions and outflow ram pressure. This may explain the failure of pressure equilibrium to account for the properties of cool halo gas inferred from COS observations.

In future work on this subject we will use a similar numerical setup to study the effect of filamentary accretion, magnetic fields, anisotropic conduction and viscosity, and halo growth rate.

high mass loading

low mass loading

Oblique Protoplanetary Disks

In addition to my work on the CGM and galactic winds, working with Chris McKee I made contributions to the field of protostellar disk and planet formation using a similar idealized simulation approach. The Kepler mission found many Jupiter mass planets with very short orbital periods (P<10 days). How these "hot Jupiters" migrated to such close-in orbits is debated. Observations of non-zero stellar obliquity---the angle between the host star's spin axis and the hot Jupiter's orbital axis---seemed to favor orbital angular momentum loss due to tidal dissipation from the star-planet interaction during close periastron passages induced by torques from a distant third body (Fabrycky and Tremain 2007). However, using hydrodynamic and magnetohydrodynamic simulations of protostar formation, I demonstrated that a wide range of stellar obliquities may be produced as a byproduct of forming a star within a turbulent environment (Fielding+15). I developed a simple semi-analytic model that reveals this connection between the turbulent motions and the orientation of a star and its disk. Our results are consistent with the observed obliquity distribution of hot Jupiters implying that the misaligned hot Jupiters may have instead migration due to tidal dissipation in the disk (Goldreich and Tremaine 1980). We have also applied this same concept to explain the observed misalignment of protostellar binaries outflows (Offner et al. 2016).

--- Drummond Fielding, Christopher McKee, Aristotle Socrates, Andrew J. Cunningham, and Richard I. Klein, The Turbulent Origin of Spin-Orbit Misalignment in Planetary Systems, MNRAS, 450, 3306 (2015).

--- Stella Offner, Michael Dunham, Katherine Lee, -- --Héctor Arce, and Drummond Fielding, The Turbulent Origin of Outflow and Spin Misalignment in Multiple Star Systems , ApJL, 827, 110 (2016).

Three Panel
							   Misaligned MHD Disk

The Yarkovsky (YORP) Effect

As an undergraduate I did a fun summer project with Ned Wright at UCLA. Using Ned's rotating, cratered, thermophysical astroid model I modeled the light curves of WISE observed asteroids to determine the asteroids' thermal properties. With an asteroid's specs (diameter, albedo, rotation rate, and so on) in hand I then modeled how they will evolve due to the Yarkovsky (or YORP) effect. We focused on finding out which asteroids will make it into mean-motion resonances with Jupiter—catapulting them chaotically into other parts of the solar system, maybe even onto a collison course with the earth! We also made some cool estimates about the parent asteroid family that may have spawed the dinosaur killing asteroid.

The Yarkovsky effect is an elegant phenomenon in which the non-heliocentric component of the radiation pressure of an asteroid can torque the asteroid causing it to slowly drift. If an asteroid isn't rotating then the hottest part of the asteroid, the part emitting most of the radiation, is point back to the sun, so there is no torque. But, if an asteroid is rotating then due to thermal inertia the hottest part is offset from the sub-solar point (this is the same reason why the hottest time of day isn't noon but 1 or 2 pm), which means the asteroid is radiating away from the sun, and so there is a torque. The direction of the torque depends on the direction the asteroid is rotating.

About Me

I have moved! I am now a Flatiron Research Fellow at the Center for Computational Astrophysics in the Flatiron Institute in New York City. You can find my new personal webpage here


dfielding ‘at’ flatironinstitute ‘dot’ org

162 5th Ave.
Flatiron Institute
Center for Computational Astrophysics
New York, NY 11206


Please go to my new homepage to find an up to date CV, you can find it here
A list of my publications can be found here.

There are no search results for "{{query}}"