1 Enrique PérezEduardo LacerdaRosa González Delgado Rubén García-Benito Sebastián F. Sánchez Bernd Husemann André Amorim Rafael López Natalia Vale Asari Clara Cortijo Resolving galaxies in x,y,t,Z Methods, uncertainties & results

… papers so far …

3 From datacubes to: (R) R [HLR] Radial profiles log t [yr] SFR(R,t) R [HLR] Maps of mean stellar age, Z, SFR, A V,…

4 Output spectra corrected for: redshift Galactic extinction etc … STARLIGHTSTARLIGHT wavelength CALIFA spectral cube spatial binning The PyCASSO pipeline Python CAlifa Starlight Synthesis Organizer M *, v *,  *, A V,,, SFH,... as a function of (x,y) !!

5 + Spectral masks Bad pixel flags Correlated errors Calibration issues... Spatial masks Voronoi zones (“segmentation maps”) PyCASSO: Some “technical details” S/N ~ 20

6 Decomposing galaxy spectra with STARLIGHT L gal    M SSP (t,Z)  x SSP( ;t,Z) x e -   SFH: mass or light fractions  Pop vector Spectral Base SSPs from BC03, “CB07”,... Granada + Vazdekis  M 1 + M 2 + M Observables Full spectrum: F Dust … t,Z

Conroy 2013 (Inverse) Spectral Synthesis DERIVE Star Formation History

8 Asari 07 SFH DownsizingDownsizing SF-galaxies   SSP( |t,Z) log t [yr]

Example spectral fits: Nucleus R = 1 HLR  Spectral residuals of 1 – 4%.  Residuals improved significantly w/new reduction pipeline

10 CALIFA 277 PyCASSO  2D maps of A V, mass, mean ages, Z’s, kinematics, SFRs …

11 … + Mass assembly history in 2D… log t [yr] CALIFA 277

12 … + 1D radial & age profiles R / HLR (R) Radius [HLR] (R) A V (R) SFR (R) Light growth(t)Mass growth(t) SFR (t) % Light (t)% Mass(t) Mass (R)Light(R) You-name-it (t,R) log t [yr] CALIFA 277

13 … + Radius x Age: azymuthally averaged SFHs Radius [HLR] log t [yr] CALIFA 277

nucleus < 0.5 HLR < HLR > HLR Mass build up as a function of R MrMr log t(80%)

nucleus < 0.5 HLR < HLR > HLR Mass build up as a function of R log t(80%) MrMr

nucleus < 0.5 HLR < HLR > HLR Mass build up as a function of R log t(80%) MrMr

nucleus < 0.5 HLR < HLR > HLR Mass build up as a function of R log t(80%) MrMr

nucleus < 0.5 HLR < HLR > HLR Mass build up as a function of R log t(80%) MrMr

nucleus < 0.5 HLR < HLR > HLR Mass build up as a function of R log t(80%) MrMr

...  Noise & shape produce uncertainties of 0.10 – 0.15 dex in global properties like mean ages, mean Z’s, stellar mass surface density.  Shape errors are mostly absorbed by A V.  Uncertainties due to choice of SSP models are larger (~ x 2)  Monte Carlo simulations: noise & shape perturbations  Fits with 3 different sets of evolutionary synthesis models  Investigate spectral residuals; compare CALIFA x SDSS

Noise & shape simulations: input x output

22 The main ingredient: SSP( ;t,Z)  evolutionary synthesis Bruzual & Charlot 03 Gonzalez Delgado 05 Vazdekis 10 Le Borgne 04 …

Galaxy- averaged properties. ( 107 points ) Experiments w/diff SSPs: GM x CB spaxels from 107 galaxies

Galaxy- averaged properties. ( 107 points ) Experiments w/diff SSPs: BC x CB spaxels from 107 galaxies

Experiments w/diff SSPs: GM x CB x BC

PyCASSO: A powerful tool to digest datacubes Galaxies grow inside out Mass builds up faster for more massive galaxies, at any R! Downsizing = Downsizing(R) Relative inner-outer age difference M * = 7 x M o Theory says this is ~ where AGN and SN (low M) feedback are minimal… Galaxies are 20% smaller in mass than in light 3/4 due to  age and 1/4 due to  A V Spatially averaged and integrated properties correlate very well. Both match the properties at R = 1 HLR Effective radii are more effective than you may have thought! The local stellar mass density drives the SFH is disks, but in bulge dominated systems the total stellar mass is a more fundamental property Summary