ScienceProblem: MultiVariateGalaxyProperties

PrimaryActor:

Research astronomer

---

ScienceGoal:

To provide a means of testing/optimising semi-analytical galaxy formation models in detail, using multivariate galaxy properties.

---

DataSets:

Any large, multi-band photometric catalogue - e.g. the federation of the optical data from the Sloan Digital Sky Survey and the near-infrared data from UKIDSS.

---

ProblemDescription:

Elaborate models now exist (e.g. Cole et al. 2000) which seek to relate the observed properties of galaxies to basic physical properties and cosmology. These so-called semi-analytic galaxy formation models combine a method of tracing the evolution of the cosmological density field, and of dark matter halos contained therein, with a series of prescriptions that relate the basic physical properties of these model halos to observable properties of galaxies.

These prescriptions seek to encapsulate existing knowledge of the physics of galaxies, but they necessarily include a number of parameters which can only be set by matching observational constraints, so the optimisation of the model involves the comparison of model predictions under a range of parameter choices with a series of observational datasets. Typically, these are distributions of a particular observable quantity over a galaxy sample - e.g. the luminosity function of galaxies in the B band and the K band, the distribution of scale-lengths of spiral galaxies, the Tully-Fisher relation - and the success of the model is judged by how well the model predicts the distributions of a further set of observable quantities, such as the colour-magnitude relation for elliptical galaxies, the galaxy colour-morphology relation (i.e. B-V colour against bulge/disk ratio).

---

CurrentSolution:

Currently, these distributions are treated independently, both when used as constraints to fix model parameters and when used to judge the success of the model through comparison with further observational data. In the future, it would be preferable if these properties could be studied simultaneously. If one thinks of the galaxies populating a multi-dimensional space, where each axis refers to a different galaxy observable, then the current procedure involves studying a series of projections of this distribution onto different sets of axes, which is clearly not using all the information contained in the full distribution of particles in the multi-dimensional space. Not only might this be helpeful for the optimisation of the model parameters, but it certainly give greater power for falsifying the resultant model if one were able to use the full information available in the galaxy database.

---

VOSolution:

To do this requires a significant computational resource, and the integration of simulated and observed datasets. It is likely that it will be easier to generate the simulated data on a computer on the same local network as the galaxy catalogue, than it is either to move a copy of the catalogue or the full simulation dataset to the other, so this is another example of a VO problem requiring the upload of code onto a data centre's computer system. An additional concern here is that generating and storing the simulated datasets may need significant computational resources, especially if several iterations are required to optimise the model, and, then, if more computations are required to generate predictions from the model for further observables on the basis of which the success of the model can be judged.

---

KeyReferences:

Cole S., Lacey C.G., Baugh C.M., Frenk C.S., 2000, MNAS, 319, 168

---

---

GoodStyle: Please add comments below. This area should be used for refinement of the above document. If you want to ask questions or start a dialogue with the author, please use (or create) a topic in the Science Problems Forum. For other ScienceProblems, refer to the ScienceProblemList.
Author: Once the refinements here and comments in the forum die down, perhaps you could rewrite the problem, incorporating the comments and refinements.

---

Similar to VO.GalaxyMorphologyProbingStarFormation but this could be done now as the data exists at this stage within AG sites.

-- NicholasWalton - 17 Apr 2002

---

-- BobMann - 09 Feb 2002