Observational evidence for accelerating universe

[Gort]

This question was triggered by the fact that Adam Riess is making his lecture rounds at a local University. So, wanting to be prepared, I pulled out his old Nobel lecture, which nicely described the techniques used for the high-z s/n measurements. I was particularly interested in how they estimated the s/n "real" apparent magnitude, applying correction factors such as K-correction, host galaxy dust extinction, Milky Way dust extinction, and others. He emphasized, in his lecture, the need to subtract host galaxy light from the s/n itself, so that the s/n brightness is not over-estimated. He ended up using some custom software developed by a grad student, Brian Schmidt. All well and good, but he didn't explain how it was calibrated. It occurred to me that if his software preferentially subtracted too much light from dimmer galactic images, the s/n would appear "too faint" relative to what would be expected from a non-accelerating universe. An erroneous "acceleration" could be concluded. But there were other teams who must have encountered a similar problem. Surely they didn't use the same software to subtract the host galaxy's light. But the technique still must be calibrated (to eliminate apparent magnitude-based bias). I couldn't find any details in the literature. Anybody know how this is done?

 

[Chalnoth]

My naive guess would be that they could perform the calibration by observing nearby locations that don't have any galaxies visible along the line of sight. It would have to be near to the supernova because the brightness of the dust in our galaxy will vary over the sky.

 

[Gort]

Although possible, I think that would be difficult. They would have to try and measure "apparent magnitude bias" by comparing the expected apparent magnitude from distances measured by other means (such as Cepheids). Not sure the Cepheid accuracy is sufficient to bring out the bias. My feeling is that it would be in the noise. The bias would only become statistically significant with dim (high-z) objects. But that's also my naive guess. I can always ask Dr. Riess when he's in town!