Forecasting local primordial non-Gaussianities from UNIONS Lyman-break galaxies and <i>Planck</i> CMB lensing

Context. Local primordial non-Gaussianities, characterized by the parameter f NL loc , provide a powerful window into the physics of inflation. Cross-correlating high-redshift tracer samples with the Cosmic Microwave Background (CMB) lensing potential offers a particularly robust probe of f NL loc , mitigating imaging systematics that typically affect large-scale measurements from tracer auto-spectra. In this context, the Ultraviolet Near Infrared Optical Northern Survey ( UNIONS ) enables the selection of u -dropout high-redshift Lyman-break galaxies (LBGs). Aims. We aim to forecast the expected precision on f NL loc that is achievable from analyzing the cross-correlation power spectrum between the distribution of UNIONS -selected LBGs and the CMB lensing potential measured by the Planck satellite. Methods. We performed a Markov chain Monte Carlo forecast to estimate the uncertainties on f NL loc and on a galaxy bias parameter b 0 , which captured our uncertainty in the tracer bias. Results. We forecast σ ( f NL loc ) = 34 for an idealized photometric sample of r &lt; 24.3 LBGs selected with a Random Forest classification algorithm from UNIONS -like ugriz imaging, with a resulting surface density of 1100 deg −2 over 3730 deg 2 . This precision can be improved to σ ( f NL loc ) = 20 after spectroscopic follow-up with the Dark Energy Spectroscopic Instrument ( DESI ), during its next phase starting in 2029, DESI-II . We also tested a more realistic selection using early UNIONS data, based on a u -dropout color cut over the ugr imaging, which yields a denser sample of r &lt; 24.2 objects at 1400 deg −2 over 4760 deg 2 . From this sample–covering a larger footprint and expected to have a higher large-scale galaxy bias–we forecast an improved constraint of σ ( f NL loc ) = 20, with a similar precision that is achievable after DESI-II follow-up. In addition, we performed a preliminary validation of the redshift distribution using the clustering-redshift method with DESI DR1 data, confirming the calibration from deep, small-area photometric fields. However, accounting for uncertainties in the clustering-redshift distribution significantly degrades the f NL loc ) constraining power.