The Effect of Turbulence on Nebular Emission Line Ratios

William J. Gray, Evan Scannapieco

Research output: Contribution to journalArticlepeer-review

17 Scopus citations


Motivated by the observed differences in the nebular emission of nearby and high redshift galaxies, we carry out a set of direct numerical simulations of turbulent astrophysical media exposed to a UV background. The simulations assume a metallicity of Z/Z= 0.5 and explicitly track ionization, recombination, charge transfer, and ion-by-ion radiative cooling for several astrophysically important elements. Each model is run to a global steady state that depends on the ionization parameter U, and the one-dimensional turbulent velocity dispersion, σ1D, and the turbulent driving scale. We carry out a suite of models with a T = 42,000 K blackbody spectrum, ne = 100 cm-3, σ1D and ranging between 0.7 and 42 km corresponding to turbulent Mach numbers varying between 0.05 and 2.6. We report our results as several nebular diagnostic diagrams and compare them to observations of star-forming galaxies at a redshift of z ≈ 2.5, whose higher surface densities may also lead to more turbulent interstellar media. We find that subsonic, transsonic turbulence, and turbulence driven on scales of 1 parsec or greater, have little or no effect on the line ratios. Supersonic, small-scale turbulence, on the other hand, generally increases the computed line emission. In fact with a driving scale ≈ 0.1 pc, a moderate amount of turbulence, σ1D = 21-28 km can reproduce many of the differences between high and low redshift observations without resorting to harder spectral shapes.

Original languageEnglish (US)
Article number132
JournalAstrophysical Journal
Issue number2
StatePublished - Nov 10 2017


  • ISM: abundances
  • ISM: atoms
  • astrochemistry
  • turbulence

ASJC Scopus subject areas

  • Astronomy and Astrophysics
  • Space and Planetary Science


Dive into the research topics of 'The Effect of Turbulence on Nebular Emission Line Ratios'. Together they form a unique fingerprint.

Cite this