Algunas ideas respecto al GTMex

Download Report

Transcript Algunas ideas respecto al GTMex

PGos3: the prototype version of the Legacy Tool
Roberto Terlevich
Elena Terlevich INAOE, Mexico Coordination
INAOE, Puebla, Mexico and
IoA, Cambridge, UK
Miguel Cerviño IAA, Spain
Coordination
Peter Anders Gottingen University, Germany
Emanuele Bertone INAOE, Mexico
SSP & galaxy modeling
Atmosphere Models
Sandro Bressan Observatorio di Padova, Italy
SSP & galaxy modelling
Uta Fritze-v.Alvensleben Gottingen University, Germany
Jesús González INAOE, Mexico
La Thuile,
March 8th 2005
SSP & galaxy modeling
Observational Databases, Data Mining
Valentina Luridiana IAA, Spain
Photoionization modelling
Conceptual design, Documentation,
Ángeles López INAOE, Mexico
System support
Aurelio López INAOE, Mexico
Conceptual design
Miguel Martínez INAOE, Mexico
System support
Christophe Morisset UNAM, Mexico System development, VOTable management,
Photoionization models
Enrique Pérez IAA, Spain
Conceptual design
Carlos Rodrigo LAEFF, Spain
management, VOTable management
System development, PGos3 design and database
Roberto Terlevich INAOE, Mexico
Conceptual design, Observational Databases
The need for a theory-data intelligent interface
●
●
●
●
●
Many of the most fundamental problems in star cluster formation and evolution
seem to be on the verge of solution by combining multi-wavelength observations
with theoretical modelling that integrates stellar evolution, stellar dynamics, gas
dynamics and chemical evolution, as well as the microphysics implicit in photo
ionisation, radiative transfer, diffusion, dust formation and mixing, to name a few.
Most groups so far have concentrated in either producing large quantities of data
or model computing. The limitation in the approach means that only a small part
of the phase space volume has been explored looking for solutions when
comparing theory and observations.
Substantial progress rests with methods that can deal simultaneously with large
observational and theoretical datasets, i.e. megadatasets.
Progress will therefore depends on improving not only the theory and
observations but the interface between them. This is the area that we are actively
exploring.
As a first stage we are developing multiwavelenght Evolutionary Synthesis
capability for megadatasets (3D spectroscopy).
Evolutionary Synthesis
●
●
●
●
Is a powerful tool to interpret the light of a stellar population
(introduced by Tinsley 1968)
Makes estimates of the history of star formation, chemical evolution
and IMF
During the last decade models have improved significantly using
the latest evolutionary tracks and stellar libraries (Cervino & MasHesse 1994;Fioc & Rocca-Volmerange 1997;Leitherer et al.
1999;Vazdekis 1999;Buzzoni 2002;Sculz et al. 2002;Bruzual &
Charlot 2003)
However it is well known that the power of this method is
extremely dependent on the reliability of the stellar models used as
ingredients (Leitherer, Fritze-v.Alvensleben & Huchra 1996).
About the Legacy Tool
●
●
●
●
We are building a sophisticated open-source software tool (the
``Legacy Tool'' ) that incorporates the most advanced theoretical highresolution synthesis models of stellar populations, chemical evolution,
abundance determination and chemo-dynamical models. The authors
of the codes belong to this collaboration.
This tool brings many dispersed efforts into a single access mode that
will guide the user via a user-friendly interface.
Includes modules of Data Mining and Machine learning algorithms to
obtain from large databases the maximum of information in a fast and
accurate manner.
An interface is also planned, with chemical evolution, photoionization,
chemodynamical and dynamical models. In addition, the LT will
include a series of utilities for the analysis of both theoretical and
observational data, e.g. the determination of colours from `filters',
extinction correctors, etc.
PGos3 present status
●
●
●
●
●
The main problem to solve when bringing to a common
environment software that has been independently
developed is that of communication and interoperability
Moreover, the results from the models should have a format
that allows easy comparison with the data.
The solution: VOtables
A working prototype of the Legacy Tool, the PG0s3 is being
implemented in the IBM super-server of the INAOE Virtual
Observatory
http://ov.inaoep.mx/
Already allows the comparison of the different ingredients
and synthesis code results. Big surprises!!
PGos3 Present Status
NUV-Optical High Resolution @ 3Myr
NUV-Optical High Resolution @ 3Myr
S/N=5
Determination of
Stellar Population
Parameters in
Galaxies Using Active
Instance-based
Machine Learning
(Solorio et al 2005)
Very Fast!
Residuals
and Lick
indeces
PGos3 Second Stage
PGos3 Third Stage
THE BEGGINING
AIMS
●
Our group aims to provide the infrastructure to
assemble interdisciplinary teams with the breath of
expertise in computation, astrophysics, statistics,
dynamics and visualization to exploit the anticipated
massive data flow from future cosmology-related
experiments, both ground and space based.