Ay 127!!Galaxy Evolution!Galaxies Must Evolve!• Stars evolve: they are born from ISM, evolve, shed envelopes or explode, enriching the ISM, more stars are born…!• Structure evolves: density fluctuations collapse and merge in a hierarchical fashion!Galaxy Evolution!Assembly of the mass!Conversion of gas into stars and v.v., feedback processes!DM dominated!Cannot be observed directly,! but may be inferred!Easy to model, mainly! dissipationless!This is what is observed, and! where energy is generated!Dissipative, and very hard to! model!Evolution Timescales and Evidence!Timescales for galactic evolution span wide range:! ~ 100 Myr - galaxy free-fall and cooling time scales! 10 -100 Myr - lifetimes of massive stars! 10 -100 Myr - lifetime of the bright phase of a luminous! Active Galactic Nucleus (?)! Few x 100 Myr - rotation period of spiral galaxy! ~ Gyr - time required for two galaxies to merge! ~ 10 Gyr - age of the Universe!Observational evidence for evolution is found in:!• Stellar populations in the Milky Way (e.g., metallicity as!!a function of stellar age, etc.)!• Systematics of nearby galaxy properties!• Properties of distant galaxies seen at earlier epoch!Theoretical Tools and Approaches!1. Assembly of the mass: numerical modeling of structure formation. Fairly well advanced, but it is hard to treat any dissipative processes very accurately. Well constrained from cosmology (LSS formation)!2. Evolution of stellar populations: based on stellar evolution models, and fairly well understood. Lots of parameters: the stellar initial mass function, star formation history, stellar evolutionary tracks and spectra as functions of metallicity. Poorly constrained a priori.!3. Hybrid schemes, e.g., “semi-analytical” models. Use both of the above to assemble comprehensive models, but not constrained very well!Observational Tools and Approaches!• Deep imaging surveys and source counts, at wavelengths from UV to FIR!– Sources are always selected in emission, and any given band has its own selection effects and other peculiarities!– With enough bandpasses, one can estimate “photometric redshifts”, essentially very low resolution spectroscopy; may be unreliable!– Measurements of clustering provide additional information!• Deep spectroscopic redshift surveys: redshifts are usually obtained in the visible, regardless of how the sources are selected!– As a bonus, one can also estimate current star formation rates and rough chemical abundances from the spectra!• Diffuse extragalactic backgrounds: an integrated emission from all sources, regardless of the flux or surface brightness limits!– Extremely hard to do!– No redshift information!Colors of Stellar Populations:"Differences in Star Formation Histories!-22 -20 -18Mi(g - r)z = 0.1SDSS, Blanton et al. 2002 1.00.5z ~ 0 galaxies, Djorgovski 1992 The red sequence:!mostly E’s, no active star formation!The blue sequence: mostly!Sp’s, with active star formation!Stellar Population Synthesis Models!• We can synthesize predicted galaxy spectra as a function of time by assuming the following:!– Star formation rate (as a function of time)!– Initial mass function!– Libraries of stellar spectra for stars of different masses and metallicities and ages, etc.!– Stellar evolutionary tracks (isochrones)!• A simple stellar population (SSP) is the result of an instantaneous burst of star formation!• We can model more complex star formation histories by adding together multiple SSPs, parameterize star formation rate as a function of time as:!– dM/dt ~ exp (-t/τ) where t is the time since the start of star formation and τ is the star-formation time scale!Modeling Evolution of Stellar Pop’s!• Stellar evolution is relatively well understood both observationally and theoretically; the key points to remember:!– Massive stars are very hot, blue, very luminous, and have very short lives; they dominate the restframe UV light!– Thus we expect largest effects in the bluer parts of the spectrum!– But there are still some modest disagreements among the models!• Star formation histories are a key assumption:!– Ellipticals are best fit by a burst of early star formation followed by “passive evolution” where they fade and get redder with time τ ~ 1 Gyr or less!– Spirals are best fit by τ ~ 3-10 Gyr – they stay bluer and don’t fade as much!– Irregulars are best fit by constant star formation rates!What We Need!• Stellar theory predicts the evolution or (stellar tracks) or stars of a given mass. There is some variation among different theoretical models!• Observations give us libraries of stellar spectra as a function of age, mass, metallicity, etc.!• We need the initial mass function (IMF) of stars!• All of these are uncertain at very low metallicities and high stellar masses!• We have to assume some star formation rate (SFR) as a function of time. Popular choices include a sharp burst, a constant SFR, or an exponentially declining one:!!"#$%&−∝∂∂τttMexpPredicted Spectral Evolution!for a simple stellar population (SSP):!a δ-function burst with a fixed metallicity and IMF!Semi-Analytical Models Semi-analytical models claim to match observations using prescriptive methods for star formation and morphological assembly Warning: Star formation is a complex, poorly-constrained phenomenon: provides a weak test of the theory (age of stars ≠ age of structures)Observing Galaxy Evolution!• If redshifts are not available, we can do source counts as a function of limiting flux or magnitude; and colors as a function of magnitude (acting as a proxy for distance - not a great approximation)!• But you really do need redshifts, to get a true evolution in time, and disentangle the various evolution effects!• The field is split observationally:!– Unobscured star formation evolution: most of the energy emerging in the restframe UV, observed in the visible/NIR!– Obscured star formation: energy from young stars reprocessed by dust to emerge in FIR/sub-mm!– They have different limitations and selection effects!!Source Counts: The Effect of Evolution!log N (per unit area and unit flux or mag) !Å log f or magnitude Æ!Luminosity evolution!moves fainter sources(more!distant and