TY - JOUR

T1 - Euclid: Testing photometric selection of emission-line galaxy targets

AU - Euclid Collaboration

AU - Cagliari, M. S.

AU - Granett, B. R.

AU - Guzzo, L.

AU - Bethermin, M.

AU - Bolzonella, M.

AU - de la Torre, S.

AU - Monaco, P.

AU - Moresco, M.

AU - Percival, W. J.

AU - Scarlata, C.

AU - Wang, Y.

AU - Ezziati, M.

AU - Ilbert, O.

AU - Le Brun, V.

AU - Amara, A.

AU - Andreon, S.

AU - Auricchio, N.

AU - Baldi, M.

AU - Bardelli, S.

AU - Bender, R.

AU - Bodendorf, C.

AU - Branchini, E.

AU - Brescia, M.

AU - Brinchmann, J.

AU - Camera, S.

AU - Capobianco, V.

AU - Carbone, C.

AU - Carretero, J.

AU - Casas, S.

AU - Castellano, M.

AU - Cavuoti, S.

AU - Cimatti, A.

AU - Congedo, G.

AU - Conselice, C. J.

AU - Conversi, L.

AU - Copin, Y.

AU - Corcione, L.

AU - Courbin, F.

AU - Courtois, H. M.

AU - Da Silva, A.

AU - Degaudenzi, H.

AU - Di Giorgio, A. M.

AU - Dinis, J.

AU - Dubath, F.

AU - Duncan, C. A. J.

AU - Dupac, X.

AU - Dusini, S.

AU - Ealet, A.

AU - Farina, M.

AU - Hook, I.

PY - 2024/9/30

Y1 - 2024/9/30

N2 - Multi-object spectroscopic galaxy surveys typically make use of photometric and colour criteria to select their targets. That is not the case of Euclid, which will use the NISP slitless spectrograph to record spectra for every source over its field of view. Slitless spectroscopy has the advantage of avoiding defining a priori a specific galaxy sample, but at the price of making the selection function harder to quantify. In its Wide Survey, Euclid was designed to build robust statistical samples of emission-line galaxies with fluxes brighter than 2 × 10−16 erg s−1 cm−2, using the Hα-[N II] complex to measure redshifts within the range [0.9, 1.8]. Given the expected signal-to-noise ratio of NISP spectra, at such faint fluxes a significant contamination by incorrectly measured redshifts is expected, either due to misidentification of other emission lines, or to noise fluctuations mistaken as such, with the consequence of reducing the purity of the final samples. This can be significantly ameliorated by exploiting the extensive Euclid photometric information to identify emission-line galaxies over the redshift range of interest. Beyond classical multi-band selections in colour space, machine learning techniques provide novel tools to perform this task. Here, we compare and quantify the performance of six such classification algorithms in achieving this goal. We consider the case when only the Euclid photometric and morphological measurements are used, and when these are supplemented by the extensive set of ancillary ground-based photometric data, which are part of the overall Euclid scientific strategy to perform lensing tomography. The classifiers are trained and tested on two mock galaxy samples, the EL-COSMOS and Euclid Flagship2 catalogues. The best performance is obtained from either a dense neural network or a support vector classifier, with comparable results in terms of the adopted metrics. When training on Euclid on-board photometry alone, these are able to remove 87% of the sources that are fainter than the nominal flux limit or lie outside the 0.9 < z < 1.8 redshift range, a figure that increases to 97% when ground-based photometry is included. These results show how by using the photometric information available to Euclid it will be possible to efficiently identify and discard spurious interlopers, allowing us to build robust spectroscopic samples for cosmological investigations.

AB - Multi-object spectroscopic galaxy surveys typically make use of photometric and colour criteria to select their targets. That is not the case of Euclid, which will use the NISP slitless spectrograph to record spectra for every source over its field of view. Slitless spectroscopy has the advantage of avoiding defining a priori a specific galaxy sample, but at the price of making the selection function harder to quantify. In its Wide Survey, Euclid was designed to build robust statistical samples of emission-line galaxies with fluxes brighter than 2 × 10−16 erg s−1 cm−2, using the Hα-[N II] complex to measure redshifts within the range [0.9, 1.8]. Given the expected signal-to-noise ratio of NISP spectra, at such faint fluxes a significant contamination by incorrectly measured redshifts is expected, either due to misidentification of other emission lines, or to noise fluctuations mistaken as such, with the consequence of reducing the purity of the final samples. This can be significantly ameliorated by exploiting the extensive Euclid photometric information to identify emission-line galaxies over the redshift range of interest. Beyond classical multi-band selections in colour space, machine learning techniques provide novel tools to perform this task. Here, we compare and quantify the performance of six such classification algorithms in achieving this goal. We consider the case when only the Euclid photometric and morphological measurements are used, and when these are supplemented by the extensive set of ancillary ground-based photometric data, which are part of the overall Euclid scientific strategy to perform lensing tomography. The classifiers are trained and tested on two mock galaxy samples, the EL-COSMOS and Euclid Flagship2 catalogues. The best performance is obtained from either a dense neural network or a support vector classifier, with comparable results in terms of the adopted metrics. When training on Euclid on-board photometry alone, these are able to remove 87% of the sources that are fainter than the nominal flux limit or lie outside the 0.9 < z < 1.8 redshift range, a figure that increases to 97% when ground-based photometry is included. These results show how by using the photometric information available to Euclid it will be possible to efficiently identify and discard spurious interlopers, allowing us to build robust spectroscopic samples for cosmological investigations.

KW - Astrophysics - Cosmology and Nongalactic Astrophysics

U2 - 10.1051/0004-6361/202449970

DO - 10.1051/0004-6361/202449970

M3 - Journal article

VL - 689

JO - Astronomy and Astrophysics

JF - Astronomy and Astrophysics

SN - 1432-0746

M1 - A166

ER -