star database + sed models
The purpose of this database is to provide stellar photometry
and fitted models in an accessible way. Most parts are automated,
which is good for including many stars, but potentially bad when the
photometry isn't simple. Take the results with a grain of
While it is possible to view individual SED pages without
being logged in, you must be logged in to search or view samples. If
you want access, please ask.
Search will try to find whatever is entered in the box, and uses the
sesame service. Thus, if sesame is down then the search will hang. If the
target exists in sdb then a link to the SED page will be returned. If not,
import will be attempted. This is commonly successful, but can fail for
various reasons. Import is not instantaneous, so wait a minute before clicking
on the target name. If this just returns the same page, trying again will not
help. Better to drop me a line.
Results for individual stars are on SED pages. These attempt to
summarise the results of model fitting to the photometry. Normally
several models are fitted, and the results from the "best" one used in
the summary ("best" means highest evidence, but a more complex
model will nearly always fit IRS spectra better). Tabs give access to the
different models fitted. The
plots are zoomable with a mouse, and some photometry information is
available with a mouseover.
Residual plots show the significance of the difference between
the photometry and the best fit model (i.e. the values used to compute
the model likelihood). They therefore do not include any uncertainty
in the model, so should not be interpreted as the significance of any
excess over the stellar photosphere.
Links along the top of the page point to various
resources. These show the SED or fractional luminosity limits, the
photometry input file and a json file with model results (see here
for example usage
), corner plots with fitted and derived
parameters, and external links to various databases. The ellipsis has
extra links, to the files produced in the fitting, Herschel PSF
fitting results, and ALMA archive (where available).
A general issue is that systems with excesses and IRS spectra
commonly prefer fits that include more than one disk component. This
is simply because it is fairly easy to improve the fit sufficiently
that the penalty for added model complication is outweighed. The
current solution is that by default two-component fits are carried out
when an excess is present, but only considered to be the best fit when
they are explicitly configured as such (via a table that I
control). The two-component fits can be viewed in the tabs, and I will
over time (with your help) set the configurations such that each
system has the "best" fit prioritised.
Stars are grouped into samples, found at the link from the home
page. These are simply directories with files that display various
sample properties, and links to individual stars. An HR diagram and
some flux vs. radius plots are available. Some sample properties and
the fitting results can be downloaded in votable form. The photometry
table contains observed and model fluxes for all filters that sdf knows
The filters used can be visualised.
Plots used to aid and verify absolute and relative calibration
between photometric systems are available. These can take some time to