You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: whitepaper/Cosmology/supernovacosmology.tex
+17-17Lines changed: 17 additions & 17 deletions
Original file line number
Diff line number
Diff line change
@@ -67,22 +67,22 @@ \subsection{Introduction}
67
67
68
68
On the other hand, the Wide-Fast-Deep (WFD) aspect will make the LSST survey
69
69
the first to scan a very large area of the sky for SNe. SNe that are detected
70
-
and well characterized by the WFD will provide
70
+
and well characterized by the WFD will provide
71
71
\begin{itemize}
72
72
\item a large, well-calibrated low redshift sample ($z \lesssim0.1$) to replace/supplement the current set of low redshift supernovae from a mixture of surveys. Such a large, clean low redshift sample is crucial in {\emph{providing a longer lever arm for the determination of cosmological parameters from supernovae.}}
73
-
\item a low and medium redshift ($z \lesssim0.8$ and peaking at $z \sim0.4$ ) sample spanning large areas of
74
-
the sky and therefore with the ability of {\emph{tracing large scale structure}} in a novel way, particularly due to
75
-
the inclusion of estimates radial distances. This will be possible by combining redshift estimates from supernova light
76
-
curves in conjunction with photometric redshifts from host galaxies. Such a sample could also be used to probe the
73
+
\item a low and medium redshift ($z \lesssim0.8$ and peaking at $z \sim0.4$ ) sample spanning large areas of
74
+
the sky and therefore with the ability of {\emph{tracing large scale structure}} in a novel way, particularly due to
75
+
the inclusion of estimates radial distances. This will be possible by combining redshift estimates from supernova light
76
+
curves in conjunction with photometric redshifts from host galaxies. Such a sample could also be used to probe the
77
77
{\emph{isotropy of the late time universe.}}
78
-
\item This large sample of SNe will also enable further sharpening of our understanding of the properties of the SN population of both Type Ia and core-collapse SNe (see \autoref{sec:transients:SNtransients}). Aside from the science described in \autoref{sec:transients:SNtransients}, this understanding will also be extremely important to the goal of SN cosmology from LSST. When selecting supernovae satisfying specific criteria from observations in magnitude limited surveys, a lack of understanding of the population properties leads to selection biases in SNIa cosmology as well as the steps in photometric classification~\cite{2017ApJ...836...56K,2016ApJ...822L..35S}.
78
+
\item This large sample of SNe will also enable further sharpening of our understanding of the properties of the SN population of both Type Ia and core-collapse SNe (see \autoref{sec:transients:SNtransients}). Aside from the science described in \autoref{sec:transients:SNtransients}, this understanding will also be extremely important to the goal of SN cosmology from LSST. When selecting supernovae satisfying specific criteria from observations in magnitude limited surveys, a lack of understanding of the population properties leads to selection biases in SNIa cosmology as well as the steps in photometric classification~\cite{2017ApJ...836...56K,2016ApJ...822L..35S}.
79
79
The WFD SN Ia
80
80
sample will dramatically increase the size of the sample available to
81
81
train such an empirical model, as well as understand the probability of
82
82
deviations and scatter from this model. Aside from issues like
83
83
calibration which need to be addressed separately, a larger sample of
84
84
such well measured SNe is probably the only way to address `systematics'
85
-
due to deviations from the empirical model.
85
+
due to deviations from the empirical model.
86
86
\end{itemize}
87
87
%The anticipated WFD sample can
88
88
%be thought of as consisting of two components: the low-redshift sample
@@ -151,7 +151,7 @@ \subsection{Target measurements and discoveries}
151
151
Because LSST will discover significantly more SNe than can be spectroscopically confirmed,
152
152
photometric classification of supernova type from multi-band light curves is crucial. While cosmology
153
153
with a photometric SNe sample with contamination from core collapse SNe is possible (see for
\caption{Example of the cadence in the 2nd season in a WFD Field
223
223
(fieldID 309) (top-panel) and a Deep Drilling Field (fieldID 744)
@@ -321,7 +321,7 @@ \subsubsection{Steps in the PerSNMetric}
321
321
discussed in \autoref{sec:\secname:targets}.
322
322
This metric is a proxy for the potential for a supernova to be detected
323
323
during its lifetime by the set of images taken in different bands by LSST. The actual
324
-
detection of SN during LSST is likely to use more stringent criteria, leading to
324
+
detection of SN during LSST is likely to use more stringent criteria, leading to
325
325
smaller numbers of supernovae in order to deal with possibly large numbers of false positives in detection. A larger
326
326
number of images taken at a time when the supernova is bright enough increases the
327
327
probability of detection. Technically, assuming that a single detection in any of the images containing
@@ -357,12 +357,12 @@ \subsubsection{Steps in the PerSNMetric}
357
357
358
358
Separating supernovae from other detected transients is being considered in
359
359
\autoref{chp:transients}. Here we concern ourselves with problem of classifying subclasses of
360
-
SNe. Multiple techniques have been proposed to solve this problem \citep{Frieman2008,sako2008, kessler2010b,
360
+
SNe. Multiple techniques have been proposed to solve this problem \citep{Frieman2008,sako2008, kessler2010b,
361
361
ishida2012, sako2014} and it is not yet clear how the
362
362
relative success of these techniques are affected by observing strategy. Work is ongoing to use the
363
363
multifaceted, machine learning pipeline developed in \citet{Lochner2016} to compare alternative
364
364
observing strategies. As this pipeline employs a variety of different feature extraction and machine learning
365
-
techniques, it is ideal to investigate the effect of observing strategy on the supernova classification. The exact
365
+
techniques, it is ideal to investigate the effect of observing strategy on the supernova classification. The exact
366
366
metric used to determine the efficacy of the classification depends
367
367
on the exact problem at hand. For producing a general purpose, well-classified set of all types of
368
368
supernovae (for example, to study supernova population statistics), one could use the AUC metric
@@ -389,9 +389,9 @@ \subsubsection{Steps in the PerSNMetric}
389
389
with better training samples and understanding of underlying
390
390
correlations of SN Ia properties and their environments. We compute a
391
391
quality metric for each SN Ia as the ratio of the square of the
392
-
intrinsic dispersion to variance of the distance indicator from the supernova.
392
+
intrinsic dispersion to variance of the distance indicator from the supernova.
393
393
$$ QM = 0.05^2/\sigma^2_{\mu}.$$ If our sample had a perfect discovery rate, and good classification (for example if we had spectroscopic classification), the uncertainty on
394
-
cosmological parameters would be entirely due to this quality metric and would be expected to scale with the quality metric as
394
+
cosmological parameters would be entirely due to this quality metric and would be expected to scale with the quality metric as
0 commit comments