Singular Templates of Imaging Cherenkov Shower distribution (STOICS): A background estimation method for Very-High-Energy $γ$-ray observations

Abstract

Analyses of Imaging Atmospheric Cherenkov Telescope (IACT) data for extended $γ$-ray sources face the issue that the field of view does not offer sufficient regions for background estimations. In cases where the source angular size exceeds or occupies a significant part of the field of view, an independent background estimation method is necessary to carry out IACT analyses and to have a better understanding of the systematic uncertainties. The proposed new method utilizes Singular Value Decomposition to extract the low-dimension representations of the distribution of cosmic-ray events in OFF runs and uses cosmic-ray-like events in the ON runs to estimate the background of $γ$-like events. Using VERITAS archival data, we demonstrate that the new method is capable of providing reliable background modeling for observations across a wide range of observing conditions.

Figures (15)

• Figure 1: The cosmic-ray event distributions in the space of $MSCL$ and $MSCW$ collected from VERITAS data of multiple blank-sky (no known $\gamma$-ray source) observations. The space is divided into eight analysis regions, with the $\gamma$-like signal region (SR) at the lower left corner containing events that pass the $\gamma$/hadron cut and seven cosmic-ray-like regions (CRs) containing events that fail the cut. The color scale shows the count of events in each bin of the histogram.
• Figure 2: Exemplary cosmic-ray event distributions in the camera frame ($2^{\circ}\times 2^{\circ}$) in the $\gamma$-like region and in the CR-like regions in different energy ranges. There exist correlations between the background shapes in the SRs and the background shapes in the CRs. The color scale shows the count of events in each bin of a mini-map. The color scales are not the same across different mini-maps and are saturated for the bin containing the largest count in a mini-map.
• Figure 3: Exemplary singular value $\sigma_{k}$ of a profile matrix compiled of a list of $\gamma$-ray-free runs in VERITAS database.
• Figure 4: Exemplary singular vectors $\vec{v}_{k}$ with $k=0$ (top) and $k=5$ (bottom), constructed from the distributions of cosmic-ray events at the reconstructed energies. The lower-order singular vector gives the generic shape of the background function, while the higher-order vector gives more detailed features. The color scales are centered at zero but are not the same across different mini-maps; they are saturated for the bin containing the largest absolute amplitude in a mini-map.
• Figure 5: Exemplary maps of model error significance in camera frame and the significance distributions (mean $\mu$ and RMS $\sigma$) in energy range $E \in [0.25,17.8]$ TeV. The background models are using $k_{\mathrm{c}}=1$ (top-left), $k_{\mathrm{c}}=2$ (top-right), $k_{\mathrm{c}}=4$ (bottom-left), and $k_{\mathrm{c}}=32$ (bottom-right). The figure demonstrates that by including more singular vectors (higher $k_{\mathrm{c}}$), the background model achieves better accuracy.