'REFF' values are all 0.0

[jbkalmbach]

While working on Twinkles, discovered that all the values coming out of OM10 for 'REFF' are 0.0.

[drphilmarshall]

I think this is because we put in REFF as a placeholder, for some bright future day when we woudl draw lens galaxies from a realistic fundamental plane model, or something. @mbaumer, perhaps we could add lens galaxy size to something we paint on from a supplied catalog, at the same time as we do the colors?

On Thu, Oct 22, 2015 at 12:17 PM, Bryce Kalmbach notifications@github.com wrote:

While working on Twinkles, discovered that all the values coming out of OM10 for 'REFF' are 0.0.

— Reply to this email directly or view it on GitHub https://github.com/drphilmarshall/OM10/issues/33.

[drphilmarshall]

@astrosonnen - Thanks for taking a look at this at the LSST DESC Hack Day. I thought the Bernardi et al multivariate Gaussian fundamental plane model was working surprisingly well! Although as @thomascollett pointed out, Hyde and Bernardi might be an even better way to go. Let's do some comparisons. I am particularly interested to see if we can reproduce the observed FB at z=0.5 from Treu et al. The timeline here is set by Twinkles 1 Run 2: it'd be great to have realistic OM10 lens galaxy i-band sizes (and potentially better i-band magnitudes as well!) by the end of April, so that we can drop them into the Twinkles field. What do you think? Doable?

[astrosonnen]

Hi Phil,

I decided that I’m going to boycott the Fundamental Plane from now on, because having Re on both sides of an equation and then have the scatter given in terms or Re.. at fixed combination of the other axis, which includes Re.. is just infuriating. The best way to go, in my opinion, is to fit a 3d Gaussian to the distribution of Mstar, Re and sigma of the SLACS lenses and just use that. I just need to figure out how to write down the covariance matrix for said Gaussian. For the sake of better interpretation, I would still like to write this Gaussian in such a way that the stellar mass is the fundamental parameter and Reff and sigma depend on it, with coefficients that explicitly describe the dependence of Re on mstar and the dependence of sigma on mstar. In contrast, I have no idea on how to interpret the mixing coefficients in the Bernardi et al. 2003 Gaussian.

I’m gonna need to do this anyway for my weak lensing paper so I think that’s doable by the end of April.

Cheers, Ale

On 30 mar 2016, at 17:02, Phil Marshall notifications@github.com wrote:

@astrosonnen https://github.com/astrosonnen - Thanks for taking a look at this at the LSST DESC Hack Day. I thought the Bernardi et al multivariate Gaussian fundamental plane model was working surprisingly well! Although as @thomascollett https://github.com/thomascollett pointed out, Hyde and Bernardi might be an even better way to go. Let's do some comparisons. I am particularly interested to see if we can reproduce the observed FB at z=0.5 from Treu et al. The timeline here is set by Twinkles 1 Run 2: it'd be great to have realistic OM10 lens galaxy i-band sizes (and potentially better i-band magnitudes as well!) by the end of April, so that we can drop them into the Twinkles field. What do you think? Doable?

— You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub https://github.com/drphilmarshall/OM10/issues/33#issuecomment-203690690

[drphilmarshall]

Thanks @astrosonnen, this sounds like a good plan. To get the evolution right, should we use both SLACS and SL2S lenses? And so are we actually talking about a simple Gaussian approximation of SL2S Paper V ? Glad to hear that this work will kill two birds with one stone :-)

On Thu, Mar 31, 2016 at 10:48 PM, Alessandro Sonnenfeld < notifications@github.com> wrote:

Hi Phil,

I decided that I’m going to boycott the Fundamental Plane from now on, because having Re on both sides of an equation and then have the scatter given in terms or Re.. at fixed combination of the other axis, which includes Re.. is just infuriating. The best way to go, in my opinion, is to fit a 3d Gaussian to the distribution of Mstar, Re and sigma of the SLACS lenses and just use that. I just need to figure out how to write down the covariance matrix for said Gaussian. For the sake of better interpretation, I would still like to write this Gaussian in such a way that the stellar mass is the fundamental parameter and Reff and sigma depend on it, with coefficients that explicitly describe the dependence of Re on mstar and the dependence of sigma on mstar. In contrast, I have no idea on how to interpret the mixing coefficients in the Bernardi et al. 2003 Gaussian.

I’m gonna need to do this anyway for my weak lensing paper so I think that’s doable by the end of April.

Cheers, Ale

On 30 mar 2016, at 17:02, Phil Marshall notifications@github.com wrote:

@astrosonnen https://github.com/astrosonnen - Thanks for taking a look at this at the LSST DESC Hack Day. I thought the Bernardi et al multivariate Gaussian fundamental plane model was working surprisingly well! Although as @thomascollett https://github.com/thomascollett pointed out, Hyde and Bernardi might be an even better way to go. Let's do some comparisons. I am particularly interested to see if we can reproduce the observed FB at z=0.5 from Treu et al. The timeline here is set by Twinkles 1 Run 2: it'd be great to have realistic OM10 lens galaxy i-band sizes (and potentially better i-band magnitudes as well!) by the end of April, so that we can drop them into the Twinkles field. What do you think? Doable?

— You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub < https://github.com/drphilmarshall/OM10/issues/33#issuecomment-203690690>

— You are receiving this because you were assigned. Reply to this email directly or view it on GitHub https://github.com/drphilmarshall/OM10/issues/33#issuecomment-204256557