0

So far, distances in the universe, using standard candles, have been measured out to a red shift of around 3, at a co-moving distance of around 20 billion light years. It was these measurements that allowed Saul Perlmutter to deduce that the expansion of the universe is currently accelerating. If this relationship could be extended to higher red shifts this would allow for the better comparison of different models of the universe.

The interpretation of any single measurement in astronomy is subject to considerable uncertainty. Is the luminosity of a galaxy due to a small galaxy nearby or a large galaxy, far away? Is red shify cosmological red shift or due to intervening dust clouds? Is the image subject to gravitational lensing? But AI generated algorithms based on data from many millions of observations may allow for much more accurate interpretations, eg without the need for spectroscopy, which is slow and costly, particularly for very distant objects.

See https://www.researchgate.net/publication/381850802_pop-cosmos_Scaleable_inference_of_galaxy_properties_and_redshifts_with_a_data-driven_population_model?_tp=eyJjb250ZXh0Ijp7ImZpcnN0UGFnZSI6InNjaWVudGlmaWNDb250cmlidXRpb25zIiwicGFnZSI6InNjaWVudGlmaWNDb250cmlidXRpb25zIn19 for more info on the new astronomy

Qmechanic
  • 220,844
John Hobson
  • 123

1 Answers1

1

The bigger question is would trust the conclusion reached by some AI algorithm where we are not absolutely sure of the methods it has used to reached its conclusion?

Anyway, one of the problems with measuring high-z supernovae is time dilation. When surveying for supernova, one technique is to photograph a portion of the sky and then compare the photo to a photograph of the same area some time later and see if any fresh stars have appeared or changed in brightness. A supernova has a characteristic brightness curve and has to be surveyed over a long enough period to determine if the variable brightness star really has the characteristic brightness curve of a type 1A supernova. Now a nearby supernova event might be very bright for maybe a month, but due to time dilation very distant supernovae might appear to last many months or even years, so they have to be observed over a fairly lengthy periods to capture enough of the brightness time curve to see if it a match for the correct type of supernova. In order to calculate the luminosity distance we need to long enough to capture a big enough portion of the luminosity time curve to reliably determine exactly where it is on that curve. This means each individual supernova observation is a lengthy process and more time consuming at high-z.

Another factor is peculiar motion. Peculiar motion of the Earth or the space telescope making the observation in the direction of the supernova at the time of the observation, affects the apparent redshift of the supernova and has to be factored into the calculations. Yet another factor is that gravitational lensing due to large clusters close to the path of the light rays will affect the light travel time, apparent brightness and redshift. This means that supernova have to be carefully analysed on a case by case basis that tries to take all these effects into account.

Additionally the peculiar motion of the supernovae itself will alter the observed redshift and the peculiar motion of a given supernova is not always known. This will introduce some 'scatter' into the data that should hopefully average out over large data sets. Despite all the hard work required, the database of high-z supernovae observations is growing into the hundreds now and hopefully a clearer picture will emerge soon.

Is the luminosity of a galaxy due to a small galaxy nearby or a large galaxy, far away?

One way we can tell the difference is the Tully–Fisher relation. Generally speaking, galaxies are not used as standard candles at high-z. Galaxies are used as part of the measurement ladder for calibrating standard candles that are closer to home. For example we can calculate the theoretical luminosity of a type 1A supernova but its nice to check the theoretical against actual measurements of nearer supernovae and how their theoretical luminosity distances compare to the distances of their assumed host galaxies estimated by other methods.

KDP
  • 10,288