An analysis of the first set of low-redshift ( z < 0.08 ) Type Ia supernovae monitored by the Carnegie Supernova Project between 2004 and 2006 is presented .
The data consist of well-sampled , high-precision optical ( ugriBV ) and near-infrared ( NIR ; YJHK _ { s } ) light curves in a well-understood photometric system .
Methods are described for deriving light-curve parameters , and for building template light curves which are used to fit Type Ia supernova data in the ugriBVYJH bands .
The intrinsic colors at maximum light are calibrated using a subsample of supernovae assumed to have suffered little or no reddening , enabling color excesses to be estimated for the full sample .
The optical–NIR color excesses allow the properties of the reddening law in the host galaxies to be studied .
A low average value of the total-to-selective absorption coefficient , R _ { V } \approx 1.7 , is derived when using the entire sample of supernovae .
However , when the two highly reddened supernovae ( SN 2005A and SN 2006X ) in the sample are excluded , a value R _ { V } \approx 3.2 is obtained , similar to the standard value for the Galaxy .
The red colors of these two events are well matched by a model where multiple scattering of photons by circumstellar dust steepens the effective extinction law .
The absolute peak magnitudes of the supernovae are studied in all bands using a two-parameter linear fit to the decline rates and the colors at maximum light , or alternatively , the color excesses .
In both cases , similar results are obtained with dispersions in absolute magnitude of 0.12–0.16 mag , depending on the specific filter-color combination .
In contrast to the results obtained from the comparison of the color excesses , these fits of absolute magnitude give R _ { V } \approx 1–2 when the dispersion is minimized , even when the two highly reddened supernovae are excluded .
This discrepancy suggests that , beyond the “ normal ” interstellar reddening produced in the host galaxies , there is an intrinsic dispersion in the colors of Type Ia supernovae which is correlated with luminosity but independent of the decline rate .
Finally , a Hubble diagram for the best-observed subsample of supernovae is produced by combining the results of the fits of absolute magnitude versus decline rate and color excess for each filter .
The resulting scatter of 0.12 mag appears to be limited by the peculiar velocities of the host galaxies as evidenced by the strong correlation between the distance-modulus residuals observed in the individual filters .
The implication is that the actual precision of Type Ia supernovae distances is 3–4 % .