How accurate are infrared luminosities from monochromatic photometric extrapolation?

Template-based extrapolations from only one photometric band can be a cost-effective method to estimate the total infrared (IR) luminosities ($L_{\mathrm{I}\mathrm{R}}$) of galaxies. By utilizing multi-wavelength data that covers across 0.35–500,$\mu \mathrm{m}$ in GOODS-North and GOODS-South fields, we investigate the accuracy of this monochromatic extrapolated $L_{\mathrm{I}\mathrm{R}}$ based on three IR spectral energy distribution (SED) templates (\citealt[CE01]{Chary2001}; \citealt[DH02]{Dale2002}; \citealt[W08]{Wuyts2008a}) out to $z\sim 3.5$. We find that the CE01 template provides the best estimate of $L_{\mathrm{I}\mathrm{R}}$ in {\it Herschel}/PACS bands, while the DH02 template performs best in {\it Herschel}/SPIRE bands. To estimate $L_{\mathrm{I}\mathrm{R}}$, we suggest that extrapolations from the available longest wavelength PACS band based on the CE01 template can be a good estimator. Moreover, if PACS measurement is unavailable, extrapolations from SPIRE observations but based on the \cite{Dale2002} template can also provide a statistically unbiased estimate for galaxies at $z\lesssim 2$. The emission of rest-frame 10–100,$\mu \mathrm{m}$ range of IR SED can be well described by all the three templates, but only the DH02 template shows nearly unbiased estimate of the emission of the rest-frame submillimeter part.