Probable detection of HI at $z\simeq 1.3$ from DEEP2 galaxies using the GMRT

We have observed the DEEP2 galaxies using the Giant Meterwave Radio Telescope in the frequency band of 610 MHz. There are $\simeq 400$ galaxies in the redshift range $1.24 < z < 1.36$ and within the field of view $\simeq 44'$, of the GMRT dishes. We have coadded the HI 21 cm-line emissions at the locations of these DEEP2 galaxies. We apply stacking on three different data cubes: primary beam uncorrected, primary beam corrected (uniform weighing ) and primary beam corrected (optimal weighing). We obtain a peak signal strength in the range $8\hbox{--}25 \, \rm \mu$Jy/beam for a velocity width in the range $270\hbox{--} 810 \, \rm km \, sec^{-1}$. The error on the signal, computed by bootstrapping, lies in the range $2.5\hbox{--}6 \, \rm \mu$Jy/beam, implying a 2.5--4.7-$\sigma$ detection of the signal at $z \simeq 1.3$. We compare our results with N-body simulations of the signal at $z\simeq 1$ and find reasonable agreement. We also discuss the impact of residual continuum and systematics.